# Structured-data

Structured data is data organized in a predefined format, typically using a fixed schema and tabular representation.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## PriTabICL: Differentially Private Tabular In-Context Learning with Task-Aligned Cluster-Wise Synthesis

DevFeed: [PriTabICL: Differentially Private Tabular In-Context Learning with Task-Aligned Cluster-Wise Synthesis](<https://devfeed.tech/articles/pritabicl-differentially-private-tabular-in-context-learning-with-task-aligned-cluster-wise-synthesis-145006.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25073>)

Author: Syed Ahzaz Shah, Evan Bramblett, H\\'eber H. Arcolezi, Chen Gong, Qianqian Tong, Xiaochen Li

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [clusters](<https://devfeed.tech/tags/clusters.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [differential-privacy](<https://devfeed.tech/tags/differential-privacy.md>), [feature](<https://devfeed.tech/tags/feature.md>), [in-context-learning](<https://devfeed.tech/tags/in-context-learning.md>), [large-language-models-llms](<https://devfeed.tech/tags/large-language-models-llms.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [privacy](<https://devfeed.tech/tags/privacy.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [synthesis](<https://devfeed.tech/tags/synthesis.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>)

### AI overview

PriTabICL aims to preserve prediction-relevant relationships in differentially private Tabular In-Context Learning demonstrations. It partitions records into differentially private clusters, synthesizes each cluster independently using a fixed feature-target workload, and routes each query to a cluster whose synthetic records supply demonstrations without additional privacy loss for the original records. The authors report experiments on four real-world datasets; PriTabICL-PrivSyn improves on global PrivSyn in 8 of 12 settings and outperforms existing private TabICL approaches in nearly all evaluated settings. These are abstract-level reported results; no metrics or detailed experimental conditions are supplied here.

### Source excerpt

arXiv:2611.25073v1 Announce Type: cross Abstract: Tabular in-context learning (TabICL) enables large language models (LLMs) to predict target records using a handful of labeled demonstrations in their prompts. However, these demonstrations often contain sensitive personal information. Differential privacy (DP) offers principled protection, but noise-based methods introduce substantial noise, distorting the feature-label relationships essential for accurate prediction, especially as data dimensionality grows. In this work, we study DP data synthesis for generating privacy-preserving TabICL demonstrations. We identify two limitations of directly applying general-purpose DP synthesis to entire datasets: modeling a global distribution can obscure local structure and minority patterns in heterogeneous data, while measuring broad attribute dependencies under a limited privacy budget can reduce the accuracy available for relationships most relevant to prediction. More broadly, global statistical fidelity may fail to preserve heterogeneous structures needed by downstream predictive and generative models. To address these challenges, we propose PriTabICL, a general framework compatible with existing DP synthesis algorithms. It partitions records into DP clusters and synthesizes each independently, using a fixed feature-target workload to prioritize prediction-relevant statistical relationships. At inference, each query is routed to a DP cluster, and demonstrations are sampled from its synthetic dataset without additional privacy loss for the original records. Experiments on four real-world datasets demonstrate the effectiveness of our methods across heterogeneous datasets. In particular, one of our instantiations, PriTabICL-PrivSyn, improves over global PrivSyn in 8 out of 12 settings and outperforms existing private TabICL approaches in nearly all evaluated settings. Open-source code is available.

## How to Generate Synthetic Healthcare Data Using Synthea: A Hands On Tutorial

DevFeed: [How to Generate Synthetic Healthcare Data Using Synthea: A Hands On Tutorial](<https://devfeed.tech/articles/how-to-generate-synthetic-healthcare-data-using-synthea-a-hands-on-tutorial-144721.md>)

Original publisher: [Read original article](<https://www.freecodecamp.org/news/how-to-generate-synthetic-healthcare-data-using-synthea-a-hands-on-tutorial/>)

Author: Lakshmi Mahabaleshwara

Published: 2026-11-19T22:00:42Z

Content type: tutorial

Language: en

Sources: [Free Code Camp](<https://devfeed.tech/sources/freecodecamp-programming-tutorials-python-javascript-git-more.md>)

Topics: [Tutorial](<https://devfeed.tech/topics/tutorial.md>), [Data science and engineering](<https://devfeed.tech/topics/data-science-and-engineering.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Data pipelines](<https://devfeed.tech/topics/data-pipelines.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [reproducibility](<https://devfeed.tech/topics/reproducibility.md>), [data](<https://devfeed.tech/topics/data.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-pipeline](<https://devfeed.tech/tags/analytics-pipeline.md>), [applications](<https://devfeed.tech/tags/applications.md>), [csv](<https://devfeed.tech/tags/csv.md>), [data-engineering](<https://devfeed.tech/tags/data-engineering.md>), [dataengineering](<https://devfeed.tech/tags/dataengineering.md>), [generate](<https://devfeed.tech/tags/generate.md>), [healthcare](<https://devfeed.tech/tags/healthcare.md>), [healthcareai](<https://devfeed.tech/tags/healthcareai.md>), [healthcaredata](<https://devfeed.tech/tags/healthcaredata.md>), [healthtech](<https://devfeed.tech/tags/healthtech.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>), [synthetichealthcaredata](<https://devfeed.tech/tags/synthetichealthcaredata.md>), [tutorial](<https://devfeed.tech/tags/tutorial.md>), [using](<https://devfeed.tech/tags/using.md>)

### AI overview

Synthea simulates longitudinal patient records, making it useful for prototyping and testing healthcare data workflows without using real patient histories. The tutorial shows how to build it with Java 17 or newer, generate about 100 patients, enable CSV export explicitly, and explore linked patient, encounter, condition, observation, medication, and procedure files with Python and Pandas. A fixed seed helps reproduce runs; version, configuration, and reference date also affect results. Default population overflow can produce more records than requested (115 in the described run); disabling it targets the requested count but may change how closely survivor demographics match population targets. The tutorial demonstrates basic relationship and missing-value checks and reconstructing a patient's timeline. Synthetic data is not necessarily representative, clinically accurate, unbiased, or suitable evidence for real-world model performance, so validate concepts, data quality, and representativeness before use.

### Source excerpt

Healthcare generates huge amounts of data every day, from patient visits and lab results to diagnoses, medications, and medical imaging. For developers and data engineers building healthcare applicati

## 6 Sneaky Bugs That Were Hurting My Next.js Site's SEO

DevFeed: [6 Sneaky Bugs That Were Hurting My Next.js Site's SEO](<https://devfeed.tech/articles/6-sneaky-bugs-that-were-hurting-my-next-js-site-s-seo-140885.md>)

Original publisher: [Read original article](<https://hackernoon.com/6-sneaky-bugs-that-were-hurting-my-nextjs-sites-seo?source=rss>)

Author: Mr Sanzid

Published: 2026-11-18T14:59:29Z

Content type: article

Language: en

Sources: [HackerNoon](<https://devfeed.tech/sources/hackernoon.md>)

Topics: [Search engine optimization (SEO)](<https://devfeed.tech/topics/seo.md>), [Google Search](<https://devfeed.tech/topics/google-search.md>), [bugfixes](<https://devfeed.tech/topics/bugfixes.md>), [data type](<https://devfeed.tech/topics/data-type.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [2026](<https://devfeed.tech/tags/2026.md>), [3](<https://devfeed.tech/tags/3.md>), [added](<https://devfeed.tech/tags/added.md>), [admin](<https://devfeed.tech/tags/admin.md>), [after](<https://devfeed.tech/tags/after.md>), [aggregate](<https://devfeed.tech/tags/aggregate.md>), [aggregation](<https://devfeed.tech/tags/aggregation.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [api](<https://devfeed.tech/tags/api.md>), [bugs](<https://devfeed.tech/tags/bugs.md>), [cloudinary-image-urls](<https://devfeed.tech/tags/cloudinary-image-urls.md>), [data](<https://devfeed.tech/tags/data.md>), [debugging](<https://devfeed.tech/tags/debugging.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [images](<https://devfeed.tech/tags/images.md>), [mongodb](<https://devfeed.tech/tags/mongodb.md>), [mongoose-schema-validation](<https://devfeed.tech/tags/mongoose-schema-validation.md>), [next-js](<https://devfeed.tech/tags/next-js.md>), [nextjs](<https://devfeed.tech/tags/nextjs.md>), [nextjs-image-optimization](<https://devfeed.tech/tags/nextjs-image-optimization.md>), [site](<https://devfeed.tech/tags/site.md>), [structured](<https://devfeed.tech/tags/structured.md>), [structured-data](<https://devfeed.tech/tags/structured-data.md>), [technical-seo](<https://devfeed.tech/tags/technical-seo.md>), [web-development](<https://devfeed.tech/tags/web-development.md>)

### AI overview

A Next.js phone comparison site had several silent data and rendering bugs despite clean builds and logs. A one-image URL was split at commas embedded in Cloudinary URLs; the fix splits only at recognized URL boundaries. Twenty-five release dates stored as strings sorted behind Date values in MongoDB; converting them restored newest-first order. Tables required remark-gfm, and wide tables needed a scrollable wrapper. Google still treated redirected legacy URLs as canonical until recrawling, so a temporary sitemap prompted recrawls. Product JSON-LD was omitted when a phone had only a converted price, while official US list prices could qualify as offers. Vercel's free image-optimization quota caused uncached images to return 402; a custom loader delegated Cloudinary transformations to Cloudinary. The article also recommends keeping unknown specs distinct from yes/no and auditing real rendered pages and database types. These findings came from one site and are not claimed as universal.

### Source excerpt

Six silent bugs broke images, sorting, SEO, and structured data on a Next.js site. Here's how the developer discovered and fixed them.

## The day the sky emptied: rendering the FAA outage with ClickHouse FORMAT PNG

DevFeed: [The day the sky emptied: rendering the FAA outage with ClickHouse FORMAT PNG](<https://devfeed.tech/articles/the-day-the-sky-emptied-rendering-the-faa-outage-with-clickhouse-format-png-140979.md>)

Original publisher: [Read original article](<https://clickhouse.com/blog/faa-outage-clickhouse-format-png>)

Author: Alexey Milovidov

Published: 2026-11-18T09:26:18Z

Content type: article

Language: en

Sources: [ClickHouse Blog](<https://devfeed.tech/sources/clickhouse-blog.md>)

Topics: [web graphics](<https://devfeed.tech/topics/web-graphics.md>), [Image](<https://devfeed.tech/topics/image.md>), [data transformation](<https://devfeed.tech/topics/data-transformation.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [CSV](<https://devfeed.tech/topics/csv.md>), [cURL](<https://devfeed.tech/topics/curl.md>)

Tags: [ads-b](<https://devfeed.tech/tags/ads-b.md>), [air-traffic](<https://devfeed.tech/tags/air-traffic.md>), [clickhouse](<https://devfeed.tech/tags/clickhouse.md>), [faa](<https://devfeed.tech/tags/faa.md>), [faa-outage](<https://devfeed.tech/tags/faa-outage.md>), [rendering](<https://devfeed.tech/tags/rendering.md>), [sql](<https://devfeed.tech/tags/sql.md>), [us](<https://devfeed.tech/tags/us.md>)

### AI overview

The excerpt reports that ADS-B positions show US air traffic thinning sharply after about 15:40 ET during the described November 13, 2026 FAA outage: by 20:00 ET, aircraft at cruising altitude were below one tenth of early afternoon levels, while high-altitude traffic began returning after 21:00. It explains how ClickHouse renders the day as an animated PNG directly from SQL: server-side aggregation produces per-frame pixel data, then SQL layers aircraft colors, a faint all-day trace, timeline, and text into the image. Because the public endpoint caps results at about one million rows, aggregation runs in 24 hourly queries and is rendered locally. For a still comparison, a separate query renders two panels directly on the server; its read-only public user limits canvas size. The excerpt also notes tradeoffs: ordinary animation rendering holds about 1.2 GB of frames in memory, streaming uses one frame but requires time ordering and may produce files some decoders reject, and palette conversion reduces APNG size from 150 MB to 25 MB.

### Source excerpt

Visualize the November 13 FAA outage using ADS-B flight data and ClickHouse's FORMAT PNG, rendering an entire day of US air traffic directly from SQL.

## TabDAG: Structured Evolutionary Search for Tabular Feature Engineering

DevFeed: [TabDAG: Structured Evolutionary Search for Tabular Feature Engineering](<https://devfeed.tech/articles/tabdag-structured-evolutionary-search-for-tabular-feature-engineering-141811.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22828>)

Author: Anastasia Volkova (FusionBrain Lab), Andrey Kuznetsov (FusionBrain Lab), Ivan Oseledets (AXXX), Valentin Khrulkov (FusionBrain Lab, AXXX)

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [clustering](<https://devfeed.tech/topics/clustering.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [ablation](<https://devfeed.tech/tags/ablation.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ne](<https://devfeed.tech/tags/cs-ne.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [feature-engineering](<https://devfeed.tech/tags/feature-engineering.md>), [graph](<https://devfeed.tech/tags/graph.md>), [performance](<https://devfeed.tech/tags/performance.md>), [search](<https://devfeed.tech/tags/search.md>), [structured](<https://devfeed.tech/tags/structured.md>)

### AI overview

TabDAG searches for tabular features by evolving executable dependency graphs: nodes represent row-wise or data-fitted transformations, and structured, schema-constrained edits build compositions such as interactions, encodings, and clustering or neighborhood features. In experiments on ten regression and classification datasets and three model families, evolved features improve performance over raw inputs. TabDAG and TabDAG combined with OpenFE have the best mean ranks among compared methods, though pairwise rank differences between feature-engineering methods do not survive multiple-testing correction. With the search engine, model backbone, and candidate budget held fixed, constrained graph edits produce fewer failed evaluations than free-form Python or whole-graph JSON rewrites. Structural checks prevent invalid programs from reaching evaluation, but proposals may be rejected and valid graphs can still fail at runtime.

### Source excerpt

arXiv:2611.22828v1 Announce Type: cross Abstract: Feature engineering can improve tabular prediction, but useful transformations are compositional, data-dependent, and subject to leakage constraints. We present TabDAG, a framework that combines evolutionary search with a graph representation of feature-engineering programs. Nodes implement row-wise or data-fitted transformations; edges encode dependencies among raw variables, intermediate values, and final features. The representation supports interactions, statistical encodings, clustering- and neighborhood-based features, and multi-stage compositions in a single executable program. Mutations are structured edits constrained by a parent-specific schema and checked before evaluation. In our experiments, no structurally invalid program reached evaluation, although proposals could be rejected and valid graphs could still fail at runtime. Across ten regression and classification datasets and three tabular model families, evolved features improve predictive performance over raw inputs. TabDAG and its combination with OpenFE achieve the two best mean ranks among the compared methods; pairwise rank differences between feature-engineering methods do not survive multiple-testing correction. The LLM-based searches use the same backbone and candidate-evaluation budget. A controlled ablation compares free-form Python, whole-graph JSON, and constrained graph edits with the search engine, budget, and backbone held fixed. Constrained edits yield fewer failed evaluations than either rewrite baseline. The resulting graphs expose the transformations and dependencies selected by search, enabling feature-level analysis of the observed gains.

## Externalized CPDAG Summaries Improve LLM Causal Deduction

DevFeed: [Externalized CPDAG Summaries Improve LLM Causal Deduction](<https://devfeed.tech/articles/externalized-cpdag-summaries-improve-llm-causal-deduction-137283.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.31071>)

Author: Wentao Sun, Jo\\~ao Paulo Nogueira, Dominique Verchere, Mathieu Acher, Alonso Silva

Published: 2026-11-17T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Graphs](<https://devfeed.tech/topics/graphs.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [algorithm](<https://devfeed.tech/tags/algorithm.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [discovery](<https://devfeed.tech/tags/discovery.md>), [graphs](<https://devfeed.tech/tags/graphs.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [llms](<https://devfeed.tech/tags/llms.md>), [structured](<https://devfeed.tech/tags/structured.md>), [summary](<https://devfeed.tech/tags/summary.md>)

### AI overview

On the Corr2Cause benchmark, a two-turn "Structured Thinking" method improves Qwen3.5-27B's mean F1 from 77.5 to 86.5 across three seeds: the model first records a CPDAG in a constrained schema, then checks the causal claim against it. A second turn of free-text analysis performs worse than the baseline with explicit PC-algorithm instructions. The reported gain also holds with renamed variables, paraphrased hypotheses, and other models. Tests suggest most remaining errors on a fixed subset come from constructing the graph, while reliable deduction depends on the intermediate CPDAG being trustworthy.

### Source excerpt

arXiv:2609.31071v2 Announce Type: replace Abstract: Observational data rarely identify a single causal structure. Such structures are typically modeled as directed acyclic graphs (DAGs) whose arrows point from direct causes to their effects, and observed correlations and conditional independencies usually leave a whole family of DAGs that fit the data equally well. A causal claim follows from the data only if it holds in every one of them. Classical causal discovery methods such as the Peter--Clark (PC) algorithm summarize this family as a completed partially directed acyclic graph (CPDAG): an edge is directed when every compatible graph orients it the same way, and undirected when the data leave its direction open. On the Corr2Cause benchmark, large language models (LLMs) that reason in free text often lose track of this family and decide from a few local edges. We propose \emph{Structured Thinking}, a two-turn method in which the model first summarizes the CPDAG in a constrained schema and then checks the claim against that summary. Without fine-tuning or an external solver, it raises Qwen3.5-27B from $77.5$ to $86.5$ F$_1$ (mean of three seeds) over a baseline that receives explicit PC-algorithm instructions, whereas a second turn of free-text analysis does worse than that baseline. The gain holds on test splits with renamed variables or paraphrased hypotheses and on other models. Given only the graph and the claim, the model's answer flips when the graph is replaced by a valid CPDAG under which the claim has the opposite truth value. Moreover, replacing the model's graphs with reference CPDAGs on a fixed subset shows that most remaining errors arise in building the graph rather than in reading it. Making causal structure explicit can help LLMs reason more reliably, provided the intermediate representation is itself trustworthy.

## Tabular foundation models on Databricks AI Runtime

DevFeed: [Tabular foundation models on Databricks AI Runtime](<https://devfeed.tech/articles/tabular-foundation-models-on-databricks-ai-runtime-136419.md>)

Original publisher: [Read original article](<https://www.databricks.com/blog/tabular-foundation-models-databricks-ai-runtime>)

Author: Marc Susagna Holgado; Javier Poveda Panter; Puneet Jain; Ryuta Yoshimatsu

Published: 2026-11-16T17:00:00Z

Content type: article

Language: en

Sources: [Databricks](<https://devfeed.tech/sources/databricks.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [databricks](<https://devfeed.tech/topics/databricks.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [ML model versioning](<https://devfeed.tech/topics/ml-model-versioning.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [ai](<https://devfeed.tech/tags/ai.md>), [approach](<https://devfeed.tech/tags/approach.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [attention](<https://devfeed.tech/tags/attention.md>), [batch](<https://devfeed.tech/tags/batch.md>), [boolean](<https://devfeed.tech/tags/boolean.md>), [churn](<https://devfeed.tech/tags/churn.md>), [classification](<https://devfeed.tech/tags/classification.md>), [data](<https://devfeed.tech/tags/data.md>), [data-science-and-ml](<https://devfeed.tech/tags/data-science-and-ml.md>), [databricks](<https://devfeed.tech/tags/databricks.md>), [databricks-ai](<https://devfeed.tech/tags/databricks-ai.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [learning](<https://devfeed.tech/tags/learning.md>), [machine](<https://devfeed.tech/tags/machine.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [models](<https://devfeed.tech/tags/models.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>)

### AI overview

Tabular foundation models (TFMs) use pretrained transformers to predict on new tables without task-specific training: labeled rows serve as context for in-context learning, while specialized tokenizers map numerical, categorical, and boolean values into a shared representation. Fine-tuning on a specific dataset is also possible. Architectures trade scale against attention cost: TabPFNv2 attends across rows and columns but scales poorly, while TabICLv1 compresses each row into four tokens to reduce attention cost; later models extend that approach. TFMs rely heavily on synthetic pretraining data, so benchmark results may not reflect proprietary workloads. Teams should evaluate accuracy and GPU inference cost, which can exceed CPU-based XGBoost costs; memory use also grows with rows and features because the training set is passed at prediction time. The excerpt presents Databricks AI Runtime as managed serverless GPU infrastructure colocated with data, with Unity Catalog governance and examples for zero-shot inference and KumoTabular fine-tuning. The supplied text is an excerpt and omits details of the referenced comparison table.

### Source excerpt

Many of the most valuable predictive machine learning applications rely on tabular data...

## Tabular Population Priors as Structured Retrieval for EHR Medication Set Prediction

DevFeed: [Tabular Population Priors as Structured Retrieval for EHR Medication Set Prediction](<https://devfeed.tech/articles/tabular-population-priors-as-structured-retrieval-for-ehr-medication-set-prediction-135057.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.17341>)

Author: Animesh Agarwal, Meysam Ghaffari, Nina Fatehi, Carlos Morato

Published: 2026-11-16T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Amazon Bedrock Knowledge Bases](<https://devfeed.tech/topics/amazon-bedrock-knowledge-bases.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>)

Tags: [aggregate](<https://devfeed.tech/tags/aggregate.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [set](<https://devfeed.tech/tags/set.md>), [structured](<https://devfeed.tech/tags/structured.md>)

### AI overview

For fixed-vocabulary EHR medication prediction, a population prior built from sparse condition-medication co-occurrences may provide more task-aligned retrieval than retrieving similar patient records. In a PACERAG pipeline, replacing only the patient retriever with ranked medication-class candidates raised F1 from 0.303 to 0.338 and Jaccard from 0.218 to 0.237 across 1,004 held-out MIMIC-IV admissions. The swap required no additional LLM calls or inference-time patient index. A non-LLM top-k baseline reached F1 0.248. These results are limited to the reported evaluation.

### Source excerpt

arXiv:2611.17341v1 Announce Type: new Abstract: Medication prediction from electronic health records is a fixed vocabulary structured prediction problem, yet recent clinical RAG systems often treat it like open ended case based reasoning: retrieve similar patient admissions and ask an LLM to infer a medication set. This retrieval object is indirect the model receives retrieved patient records while the target is a discrete set of medication class labels. We propose a more task aligned retrieval object: a tabular population prior ranked medication class candidates from a sparse condition medication co occurrence table estimated from training admissions. Using PACERAG as a fixed downstream pipeline and replacing only its patient retriever, this drop in swap improves F1 from 0.303 to 0.338 (+11.5% relative) and Jaccard from 0.218 to 0.237 on 1,004 heldout MIMIC IV admissions, with no additional LLM calls and no inference time patient index. A non LLM top k baseline reaches F1=0.248, confirming the prior is predictive before LLM integration. Aggregate population statistics can provide a more task aligned, efficient, and auditable retrieval source than patient level nearest neighbors for fixed vocabulary EHR set prediction.

## TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders

DevFeed: [TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders](<https://devfeed.tech/articles/trl-bench-standardizing-cross-paradigm-representation-level-evaluation-of-tabular-encoders-131928.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.09323>)

Author: Wei Pang, Xiangru Jian, Hehan Li, Zhixuan Yu, Alex Xue, Jinyang Li, Zhengyuan Dong, Xinjian Zhao, Hao Xu, Chao Zhang, Reynold Cheng, M. Tamer \\"Ozsu, Tianshu Yu

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Dataset development](<https://devfeed.tech/topics/dataset-development.md>), [DeepFact-Bench](<https://devfeed.tech/topics/deepfact-bench.md>)

Tags: [bench](<https://devfeed.tech/tags/bench.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [data](<https://devfeed.tech/tags/data.md>), [data-lake](<https://devfeed.tech/tags/data-lake.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [enrichment](<https://devfeed.tech/tags/enrichment.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [learning](<https://devfeed.tech/tags/learning.md>), [models](<https://devfeed.tech/tags/models.md>), [pipelines](<https://devfeed.tech/tags/pipelines.md>), [table](<https://devfeed.tech/tags/table.md>), [tables](<https://devfeed.tech/tags/tables.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [trl](<https://devfeed.tech/tags/trl.md>)

### AI overview

TRL-Bench compares tabular encoders through exported row, column, or table embeddings probed by shared lightweight heads, allowing evaluation across training paradigms under standardized downstream conditions. Its three suites cover column/table tasks, row tasks, and compositional Data-Lake Table Enrichment. The released assets include 50 OpenML tables with 123 verified targets, 16 row-pair linkage rewrites, and a 47,772-table lake derived from 1,379 parent tables. Results across 20 models and 16 tasks suggest performance is capability-specific: generic text encoders often lead when surface text matters, while tabular specialists excel when their pretraining objective fits. Within-table prediction and cross-table linkage favor different regimes; the strongest enrichment pipelines combine matched specialists, with end-to-end quality depending on how components fit together rather than their individual ranks alone.

### Source excerpt

arXiv:2606.09323v2 Announce Type: replace Abstract: Tabular encoders are usually evaluated inside task-specific end-to-end pipelines, so models from different training paradigms are difficult to compare directly even when they operate on similar tabular signals. We introduce TRL-Bench, a multi-granular tabular representation learning (TRL) benchmark that standardizes cross-paradigm representation-level evaluation: each encoder exports row-, column-, or table embeddings through its supported wrapper, and shared lightweight heads probe them across three suites: TRL-CTbench (column/table), TRL-Rbench (row), and TRL-DLTE (compositional Data-Lake Table Enrichment spanning all three granularities). To support this standardized setting, we release curated benchmark assets and task reformulations, including 50 OpenML tables with 123 verified targets, 16 row-pair linkage rewrites, and a 47,772-table DLTE lake derived from 1,379 parent tables. Across 20 models and 16 tasks, TRL-Bench shows that once downstream conditions are standardized, encoder quality is capability-specific rather than captured by a single leaderboard. In TRL-CTbench, generic text encoders often lead on tasks with strong surface-text signal, while tabular specialists win where their pretraining objective aligns with the task. In TRL-Rbench, within-table prediction and cross-table linkage favor different training regimes, with atomic linkage performance correlating strongly with the row-matching stage of DLTE pipelines. In TRL-DLTE, the strongest pipelines combine capability-matched specialists rather than reuse a single encoder, and top end-to-end quality depends on non-additive compositional fit rather than per-stage marginal rank alone. TRL-Bench provides a common protocol for measuring reusable signal in exported tabular representations under shared downstream conditions. Code: https://github.com/LOGO-CUHKSZ/TRL-Bench; data is released on Hugging Face.

## ANCHOR: Schema-Agnostic Threat Knowledge Graph Construction via Hybrid Ontology Discovery

DevFeed: [ANCHOR: Schema-Agnostic Threat Knowledge Graph Construction via Hybrid Ontology Discovery](<https://devfeed.tech/articles/anchor-schema-agnostic-threat-knowledge-graph-construction-via-hybrid-ontology-discovery-133254.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.01208>)

Author: Seonwoo Kim, Jinwoo Kim, Daegyu Kang, Daeseong Kim, Insup Lee

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [Threat Hunting & Intel](<https://devfeed.tech/topics/threat-hunting-intel.md>), [база знаний](<https://devfeed.tech/topics/tag-c51cc4bf76e5.md>), [graph-database](<https://devfeed.tech/topics/graph-database.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [schema design](<https://devfeed.tech/topics/schema-design.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>)

Tags: [adversary](<https://devfeed.tech/tags/adversary.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [anchor](<https://devfeed.tech/tags/anchor.md>), [apis](<https://devfeed.tech/tags/apis.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [attack](<https://devfeed.tech/tags/attack.md>), [based](<https://devfeed.tech/tags/based.md>), [bridges](<https://devfeed.tech/tags/bridges.md>), [challenges](<https://devfeed.tech/tags/challenges.md>), [combined](<https://devfeed.tech/tags/combined.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [construction](<https://devfeed.tech/tags/construction.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cyber](<https://devfeed.tech/tags/cyber.md>), [cyber-threat-intelligence](<https://devfeed.tech/tags/cyber-threat-intelligence.md>), [discovery](<https://devfeed.tech/tags/discovery.md>), [graph](<https://devfeed.tech/tags/graph.md>), [hybrid](<https://devfeed.tech/tags/hybrid.md>), [knowledge-graph](<https://devfeed.tech/tags/knowledge-graph.md>), [ontology](<https://devfeed.tech/tags/ontology.md>), [schema](<https://devfeed.tech/tags/schema.md>)

### AI overview

ANCHOR constructs cyber threat intelligence knowledge graphs across changing ontology schemas. It combines hybrid ontology discovery, which searches and navigates large schemas, with SHACL validation to enforce schema-compliant type assignments. The approach is intended to avoid manual pipeline reconfiguration and the scaling limits of inserting large ontologies into prompts. Experiments on UCO, STIX, and MALOnt reportedly show better ontology typing and schema compliance than existing baselines. With a local LLM, ANCHOR closely matches enterprise LLM typing performance, supporting privacy-preserving analysis of sensitive incident data. The supplied abstract does not include experiment details or quantify these gains.

### Source excerpt

arXiv:2606.01208v2 Announce Type: replace Abstract: Cyber threat intelligence (CTI) reports now serve as essential resources for capturing adversary tactics, techniques, and procedures observed in modern attack campaigns. While traditional CTI platforms reduce this intelligence to isolated indicators through fixed schemas such as STIX, ontology-based representations preserve the semantic relationships needed for structured threat analysis. However, existing approaches for ontology-aligned CTI extraction face three challenges: (i) schema-specific pipelines that require manual reconfiguration whenever the schema changes, (ii) prompt-based schema inclusion that fails to scale on large ontologies such as UCO, and (iii) reliance on enterprise LLM APIs that conflicts with privacy constraints when integrating sensitive internal incident data. In this paper, we present ANCHOR, a schema-agnostic CTI knowledge graph construction system that bridges LLMs and formal ontology schemas. At the core of ANCHOR is hybrid ontology discovery, a search-and-navigate mechanism that dynamically explores large-scale ontology schemas, combined with SHACL-based validation to enforce schema-compliant type assignments. Experimental results on the UCO, STIX, and MALOnt schemas show that ANCHOR outperforms existing baselines in ontology typing and schema compliance. In addition, ANCHOR with a local LLM closely matches enterprise LLM typing performance, enabling privacy-preserving CTI analysis with high fidelity.

## Flow-CAD: Flow Matching for Contextual Anomaly Detection and Explanation in Tabular Data

DevFeed: [Flow-CAD: Flow Matching for Contextual Anomaly Detection and Explanation in Tabular Data](<https://devfeed.tech/articles/flow-cad-flow-matching-for-contextual-anomaly-detection-and-explanation-in-tabular-data-133091.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.16184>)

Author: Zhong Li, Qi Huang, Yuhang Wang, Yang Cao, Thomas B\\"ack, Niki van Stein, Matthijs van Leeuwen, Lincen Yang

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Detection engineering](<https://devfeed.tech/topics/detection-engineering.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Data analysis](<https://devfeed.tech/topics/data-analysis.md>)

Tags: [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.md>), [attribution](<https://devfeed.tech/tags/attribution.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [code](<https://devfeed.tech/tags/code.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [flow-matching](<https://devfeed.tech/tags/flow-matching.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>)

### AI overview

Flow-CAD uses one global, amortized Flow Matching model to learn tabular behavior conditioned on context, avoiding simple parametric assumptions and per-context retraining. It scores anomalies with scale-normalized, truncated residual energy from the learned flow; the authors say this stabilizes scores, limits domination by extreme deviations, and enables feature-wise attribution. They report theoretical stability under velocity-estimation error and concentration bounds for the Monte Carlo estimator. Across 20 benchmark datasets and 13 competitors, Flow-CAD achieved the best average rank for PR AUC, ROC AUC, and P@n. These are abstract-level claims; the supplied text gives no benchmark details or numerical effect sizes.

### Source excerpt

arXiv:2611.16184v1 Announce Type: new Abstract: Contextual anomaly detection in tabular data aims to identify objects whose behavior is unexpected under their contextual conditions. In data-engineering pipelines, this task screens table rows whose measurements violate context-dependent integrity relationships. Most existing approaches rely on either shallow conditional models with limited distributional expressiveness or localized estimators that require per-context modeling, making it difficult to capture complex conditional relationships and scale to high-dimensional settings. We propose Flow-CAD, a new framework for contextual anomaly detection based on Flow Matching. Flow-CAD learns a single global and amortized conditional generative model that flexibly represents behavioral distributions conditioned on context, without assuming simple parametric forms or retraining local models. To enable robust and interpretable anomaly scoring, we introduce a scale-normalized and truncated residual energy derived from the learned conditional flow, which stabilizes anomaly scores, mitigates domination by extreme deviations, and supports feature-wise attribution. We further provide theoretical insights showing that the proposed residual-based anomaly score is stable under velocity estimation error, and that its Monte Carlo estimator satisfies concentration bounds. Experiments on 20 benchmark datasets with 13 competitors demonstrate that Flow-CAD has the best average rank for PR AUC, ROC AUC, and P@n, highlighting its effectiveness and practicality for contextual anomaly detection in tabular data. We provide the source code at https://github.com/ZhongLIFR/FlowCAD.

## One Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction

DevFeed: [One Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction](<https://devfeed.tech/articles/one-table-many-queries-benchmarking-generalist-tabular-models-across-generation-imputation-and-prediction-128644.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.13001>)

Author: Yuanrui Wang, Kaifang Li, Xuanyue Li, Ziyun Li, Yan Lu, Xingxuan Zhang, Peng Cui

Published: 2026-11-11T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [addition](<https://devfeed.tech/tags/addition.md>), [all](<https://devfeed.tech/tags/all.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [but](<https://devfeed.tech/tags/but.md>), [can](<https://devfeed.tech/tags/can.md>), [capabilities](<https://devfeed.tech/tags/capabilities.md>), [capability](<https://devfeed.tech/tags/capability.md>), [compare](<https://devfeed.tech/tags/compare.md>), [comparisons](<https://devfeed.tech/tags/comparisons.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [generalist](<https://devfeed.tech/tags/generalist.md>), [generation](<https://devfeed.tech/tags/generation.md>), [models](<https://devfeed.tech/tags/models.md>), [one](<https://devfeed.tech/tags/one.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [queries](<https://devfeed.tech/tags/queries.md>), [table](<https://devfeed.tech/tags/table.md>)

### AI overview

The One Table Benchmark compares generation, imputation, and prediction across 56 datasets, 46 methods, and 20 capability axes, then tests objective combinations across five backbones and five datasets. Imputation and prediction rankings align on numerical recovery and dependence preservation; generation connects to both through predictive utility, while distributional fidelity can rank models differently. Including a task's own objective yields the largest pooled gains. Generation benefits positively from the broadest range of auxiliary objectives, and prediction can improve imputation and synthetic-data utility, though effects vary by backbone and objectives' gains do not necessarily add. Some generalist models exceed specialist portfolios on utility, while fidelity favors specialization. Frozen foundation models perform strongly in prediction and numerical point imputation and show promising synthetic-data utility. The abstract identifies categorical reconstruction and faithful joint sampling as areas for improvement.

### Source excerpt

arXiv:2611.13001v1 Announce Type: new Abstract: Generation, imputation, and prediction query different aspects of the same tabular distribution, but are usually evaluated separately. This leaves unclear which capabilities they share, how their training objectives interact, and whether one model can perform all three well. We introduce the One Table Benchmark (\ours{}), spanning 56 datasets, 46 registered methods, and 20 capability axes under aligned data splits and explicit task contracts. We compare task ranks on aligned support, train five backbones under all seven non-empty objective combinations on five datasets with three seeds, and contrast fixed-model profiles with specialist portfolios. Rank comparisons connect imputation and prediction through numerical recovery and dependence preservation, and link generation to both through predictive utility; distributional fidelity follows different, sometimes opposing, rankings. Objective interventions show the largest pooled gains from including the recipient's own objective. Generation has the broadest coverage of positive pooled auxiliary effects, while prediction shows potential to improve imputation and synthetic-data utility, with responses that vary across backbones. Cross-addition effects reveal that objective benefits need not add and depend on the objectives already present. Fixed-model comparisons expose corresponding utility--fidelity trade-offs: some generalists surpass specialist portfolios on utility, while fidelity continues to favor specialization. Frozen foundation models show strong prediction and numerical point imputation, together with promising synthetic-data utility, highlighting their potential across all three tasks. These findings motivate unified tabular models that share useful conditional information while improving categorical reconstruction and faithful joint sampling.

## Mitra-v2: Scaling synthetic pretraining for tabular foundation models

DevFeed: [Mitra-v2: Scaling synthetic pretraining for tabular foundation models](<https://devfeed.tech/articles/mitra-v2-scaling-synthetic-pretraining-for-tabular-foundation-models-127494.md>)

Original publisher: [Read original article](<https://www.amazon.science/blog/mitra-v2-scaling-synthetic-pretraining-for-tabular-foundation-models>)

Author: Chris (Luyang) Kong

Published: 2026-11-10T15:37:39Z

Content type: article

Language: en

Sources: [Amazon Science homepage](<https://devfeed.tech/sources/amazon-science-homepage.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [few-shot](<https://devfeed.tech/topics/few-shot.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [classification](<https://devfeed.tech/tags/classification.md>), [data](<https://devfeed.tech/tags/data.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [pretraining](<https://devfeed.tech/tags/pretraining.md>), [scaling](<https://devfeed.tech/tags/scaling.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>), [synthetic-data-generation](<https://devfeed.tech/tags/synthetic-data-generation.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>), [v2](<https://devfeed.tech/tags/v2.md>)

### AI overview

Mitra-v2 improves tabular prediction by training the same Transformer architecture on synthetic tables with 10 times as many rows and roughly three times as many columns as Mitra-v1, plus a hybrid causal prior combining neural networks, decision trees, Gaussian-process functions, and time-dependent processes. At deployment, it briefly fine-tunes on each table and averages eight model copies, adding compute despite the in-context-learning setup. SVD wrappers handle very wide tables, while a hierarchical tree extends classification beyond the output layer's 10 classes. On TabArena, it leads the single-model field, improves on Mitra-v1 on 32 of 33 datasets both run natively, and completes all 51 within the one-hour limit. Fine-tuned Mitra-v2 ranks first on ScoringBench for calibrated regression ranges. The text also reports strong performance on TALENT, especially for classification above 10 classes. Synthetic pretraining uses no real or customer data.

### Source excerpt

Expanding the mix of synthetic data used to train a tabular foundation model enables it to adapt to new tables without new models and lead the single-model field on the TabArena benchmark.

## Out-of-the-Box Tabular Foundation Models for Bayesian Optimization in Materials Discovery

DevFeed: [Out-of-the-Box Tabular Foundation Models for Bayesian Optimization in Materials Discovery](<https://devfeed.tech/articles/out-of-the-box-tabular-foundation-models-for-bayesian-optimization-in-materials-discovery-125559.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.12491>)

Author: Klemens Fl\\"oge, Simon Bing, Leo Grinsztajn, Eliott Kalfon, Lennart Purucker, Noah Hollmann, Frank Hutter

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI for science](<https://devfeed.tech/topics/ai-for-science.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>)

Tags: [bayesian](<https://devfeed.tech/tags/bayesian.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [materials](<https://devfeed.tech/tags/materials.md>), [materials-discovery](<https://devfeed.tech/tags/materials-discovery.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>)

### AI overview

TabPFN and TabICL were evaluated as out-of-the-box surrogate models for Bayesian optimization on two standard materials science benchmarks, against Gaussian processes, random forests, and CatBoost. The abstract reports that both tabular foundation models were highly competitive without task-specific active-learning training or Gaussian-process kernel selection. It notes that Gaussian processes require task-dependent kernel choices and may lack capacity for complex relationships; the reported findings motivate further development of TFMs for active learning in materials discovery.

### Source excerpt

arXiv:2611.12491v1 Announce Type: new Abstract: Bayesian optimization is widely used in active learning loops for materials science. Gaussian processes serve as the de-facto surrogate model family, but require task-dependent kernel choices and can lack the modeling capacity for complex relationships. Tabular foundation models (TFMs) have recently emerged as the state-of-the-art for predictive tasks on structured data, without requiring bespoke kernels or hyperparameter tuning. We evaluate two TFMs, TabPFN and TabICL, as out-of-the-box surrogate models on two standard materials science benchmarks, comparing against a large field of baselines encompassing Gaussian processes, random forests, and CatBoost. Our findings indicate that TFMs are highly competitive surrogate models across the two benchmarks, without task-specific training for active learning or GP kernel selection. Our results motivate developing TFMs further as surrogates for active learning in materials discovery.

## RAD-TFM: Robust and Domain-Adapted Tabular Foundation Models

DevFeed: [RAD-TFM: Robust and Domain-Adapted Tabular Foundation Models](<https://devfeed.tech/articles/rad-tfm-robust-and-domain-adapted-tabular-foundation-models-125150.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.10230>)

Author: Matthew Peroni, Franck Le, Vadim Sheinin

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [adversarial](<https://devfeed.tech/tags/adversarial.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [performance](<https://devfeed.tech/tags/performance.md>), [robust](<https://devfeed.tech/tags/robust.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>), [synthetic-data](<https://devfeed.tech/tags/synthetic-data.md>), [synthetic-data-generation](<https://devfeed.tech/tags/synthetic-data-generation.md>)

### AI overview

RAD-TFM improves tabular foundation models by adversarially adapting their synthetic data generator to emphasize challenging datasets, using an optimality gap to guide training. It also aligns generated data with a target domain so the adversary produces realistic datasets. The model-agnostic pipeline uses synthetic data alone. On TabPFN V2, it improved results across six benchmarks, with up to a 10.8 percentage-point gain in mean normalized AUC over the strongest baseline, using fewer than 100,000 additional synthetic datasets (under 0.1% of the original pretraining data). Applying the pipeline to TabICL V2 also improved performance across three benchmarks. These are abstract-level results; no benchmark details or implementation steps are provided.

### Source excerpt

arXiv:2611.10230v1 Announce Type: new Abstract: The development of tabular foundation models (TFMs) has accelerated in recent years, outperforming traditional ML methods on many tabular benchmarks. A key finding is that TFMs can be pretrained entirely on synthetic datasets, opening opportunities to design data generators that encourage desirable model properties. Prior work has mainly focused on crafting high-quality priors over generators to improve overall pretraining performance. Our key insight is that parameterizing the generator distribution enables an adversarial, distributional robustness perspective: during training, we can adapt the generator to emphasize datasets that are particularly challenging for the model, which we formalize by introducing an optimality gap measure. Further, we develop a method to align the synthetic data generation with real-world datasets from a given domain, constraining the adversary to generate realistic data. Together, these algorithms comprise the Robust and Domain-Adapted Tabular Foundation Models (RAD-TFM) pipeline, a model-agnostic adversarial training framework. Applied to the TabPFN V2 classifier, RAD-TFM improves performance across 6 diverse tabular benchmarks, with up to a 10.8 percentage-point improvement in mean normalized AUC over the strongest baseline, using fewer than 100k additional synthetic training datasets, less than 0.1% of the original pretraining data. Applying the same pipeline to TabICL V2 also improves performance across three major benchmarks, demonstrating the generality of the framework. These results highlight a promising new direction for targeted adversarial continued training of TFMs using synthetic data alone.

## Cutpurse: Decomposing Explanation-Guided Model Extraction on Tabular Data

DevFeed: [Cutpurse: Decomposing Explanation-Guided Model Extraction on Tabular Data](<https://devfeed.tech/articles/cutpurse-decomposing-explanation-guided-model-extraction-on-tabular-data-122801.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.09375>)

Author: Seyma Selcan Magara, Hanru Zhang, Fatme El Zahraa Hamad, Nico Pfeifer, Mete Akg\\"un

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Cryptography](<https://devfeed.tech/topics/cryptography.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [adversary](<https://devfeed.tech/tags/adversary.md>), [api](<https://devfeed.tech/tags/api.md>), [apis](<https://devfeed.tech/tags/apis.md>), [attack](<https://devfeed.tech/tags/attack.md>), [classifiers](<https://devfeed.tech/tags/classifiers.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [lime](<https://devfeed.tech/tags/lime.md>), [model](<https://devfeed.tech/tags/model.md>), [shap](<https://devfeed.tech/tags/shap.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>)

### AI overview

Across six datasets and five target families, feature-attribution rankings gave no systematic advantage for extracting tabular classifiers. Lime's discretization edges improved fidelity by up to 41 percentage points on continuous features, but they reflect quantiles of the explainer's background data and an attacker-side grid reproduced their gains without an explanation API. Cutpurse combines a self-computed grid with diverse query generation and boundary search; when bin edges were withheld, it improved fidelity by 16.3 points on continuous-feature targets and exceeded the undefended published attack by 7.0 points. These results apply to default Lime and Shap configurations under in-domain auxiliary-input, probability-output, and same-family surrogate assumptions; they do not establish the same outcome for other explainers, target classes, or data-free attackers.

### Source excerpt

arXiv:2611.09375v1 Announce Type: new Abstract: Explanation APIs are often treated as increasing model-extraction risk, but an end-to-end attack comparison cannot show whether an explanation reveals model-specific information or merely supplies a convenient next-query coordinate. We study this distinction for explanation-guided extraction of tabular classifiers. We separate an explanation into an attribution channel, which ranks influential features, and a coordinate channel, which supplies candidate feature values for perturbing those features. We then test whether the useful bin edges are privileged to the service or reproducible from the adversary's own data. Across six datasets and five target families, using attribution rankings to choose which features to perturb provides no systematic extraction advantage. Lime's discretization bin edges can improve fidelity by up to 41 percentage points on continuous features, but those edges are quantiles of the explainer's background data rather than properties of the target model. In controlled comparisons, an attacker-side grid reproduces the useful bin-edge gains without querying an explanation API. We use these findings to derive Cutpurse, which combines a self-computed grid with diverse query generation and boundary search. Against a service that withholds bin edges, Cutpurse improves fidelity by 16.3 percentage points on continuous-feature targets. It also exceeds the undefended published attack by 7.0 percentage points. Thus, withholding default Lime bin edges alone does not prevent effective extraction under the in-domain auxiliary-input, probability-output, and same-family surrogate assumptions studied here. Our results concern default Lime and Shap configurations for tabular classifiers. They do not establish the same conclusion for other explainers, target classes, or data-free attackers.

## Bring structured and unstructured data together with Amazon Quick

DevFeed: [Bring structured and unstructured data together with Amazon Quick](<https://devfeed.tech/articles/bring-structured-and-unstructured-data-together-with-amazon-quick-115496.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/machine-learning/bring-structured-and-unstructured-data-together-with-amazon-quick/>)

Author: Puneeth Komaragiri

Published: 2026-11-03T16:49:29Z

Content type: tutorial

Language: en

Sources: [Artificial Intelligence](<https://devfeed.tech/sources/artificial-intelligence.md>)

Topics: [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [data-platforms](<https://devfeed.tech/topics/data-platforms.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Amazon S3](<https://devfeed.tech/topics/amazon-s3.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-guardrails](<https://devfeed.tech/tags/ai-guardrails.md>), [amazon-quick](<https://devfeed.tech/tags/amazon-quick.md>), [amazon-quick-suite](<https://devfeed.tech/tags/amazon-quick-suite.md>), [amazon-s3](<https://devfeed.tech/tags/amazon-s3.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analyze](<https://devfeed.tech/tags/analyze.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [artifacts](<https://devfeed.tech/tags/artifacts.md>), [aws-identity-and-access-management-iam](<https://devfeed.tech/tags/aws-identity-and-access-management-iam.md>), [aws-key-management-service-aws-kms](<https://devfeed.tech/tags/aws-key-management-service-aws-kms.md>), [aws-shared-responsibility](<https://devfeed.tech/tags/aws-shared-responsibility.md>), [best-practices](<https://devfeed.tech/tags/best-practices.md>), [custom](<https://devfeed.tech/tags/custom.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [intermediate-200](<https://devfeed.tech/tags/intermediate-200.md>), [knowledge-base](<https://devfeed.tech/tags/knowledge-base.md>), [space](<https://devfeed.tech/tags/space.md>), [structured](<https://devfeed.tech/tags/structured.md>), [technical-how-to](<https://devfeed.tech/tags/technical-how-to.md>), [unstructured-data](<https://devfeed.tech/tags/unstructured-data.md>)

### AI overview

Amazon Quick can query structured sales data and unstructured business documents through one custom chat agent. The walkthrough connects an S3 bucket to an indexed knowledge base, enriches an uploaded CSV with corrected data types and semantic guidance for text-to-SQL, and combines both sources in a Space linked to the agent. The agent can answer questions from either source or compare them in one response with citations; the example also describes generating reports and sharing the agent. Setup requires an Enterprise subscription for dataset creation, administrator-granted S3 access, and permissions to create the resources. S3 buckets must be in the same Region as Quick; document ACLs must be selected when creating the knowledge base and cannot be changed later. Shared users also need access to the Space, knowledge base, and dataset. Built-in safety screening supplements, but does not replace, access and permission controls.

### Source excerpt

Amazon Quick brings structured and unstructured data together in one experience. In this post, you connect Amazon S3 as a knowledge base, enrich a dataset, build a Space, and create a custom chat agent that answers business questions across both data types through a single conversational interface.

## LLM-Enhanced Multimodal House Price Prediction using Structured, Textual, and Visual Features

DevFeed: [LLM-Enhanced Multimodal House Price Prediction using Structured, Textual, and Visual Features](<https://devfeed.tech/articles/llm-enhanced-multimodal-house-price-prediction-using-structured-textual-and-visual-features-114575.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.01615>)

Author: Khandokar Md. Rahat Hossain, Souvik Das, Tahsina Hashem, Mohammed Eunus Ali, Tanzima Hashem

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [clustering](<https://devfeed.tech/topics/clustering.md>), [Image](<https://devfeed.tech/topics/image.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>)

Tags: [ablation](<https://devfeed.tech/tags/ablation.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [features](<https://devfeed.tech/tags/features.md>), [house](<https://devfeed.tech/tags/house.md>), [image](<https://devfeed.tech/tags/image.md>), [images](<https://devfeed.tech/tags/images.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [large-language-model](<https://devfeed.tech/tags/large-language-model.md>), [learning](<https://devfeed.tech/tags/learning.md>), [llm](<https://devfeed.tech/tags/llm.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [optimized](<https://devfeed.tech/tags/optimized.md>), [photos](<https://devfeed.tech/tags/photos.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [properties](<https://devfeed.tech/tags/properties.md>), [real-estate](<https://devfeed.tech/tags/real-estate.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [structured](<https://devfeed.tech/tags/structured.md>)

### AI overview

The proposed framework predicts house prices by combining structured property attributes with features extracted from listing text and photos. An LLM derives spatio-temporal and contextual features from descriptions; an unsupervised vision-language method groups photos into categories such as indoor, outdoor, mixed, and floorplan by aligning image clusters with pseudo-text prompts, without manual labels. The features are fused and passed to an optimized regressor. The abstract reports better performance than traditional methods and existing multimodal baselines on a large real-world dataset; ablations attribute gains to the text features and categorized visual embeddings. It provides no numerical results or dataset details.

### Source excerpt

arXiv:2611.01615v1 Announce Type: new Abstract: Accurate house price prediction is essential for real estate buyers, sellers, agents, and financial institutions. Traditional methods rely primarily on structured data, e.g., number of bedrooms, bathrooms, parkings, but they overlook rich contextual information embedded in textual descriptions and property photos. Previous deep learning approaches for house price prediction have fused multimodal data but they fall short to extract structured spatio-temporal and contextual features from texts and semantic visual information from images. In this paper, we introduce a novel multimodal framework that addresses these limitations by systematically integrating structured attributes, textual descriptions, and visual images of properties to achieve higher prediction accuracy while maintaining interpretability. We use a large language model to derive domain-specific spatio-temporal and contextual features from text, and an unsupervised vision-language approach to automatically group photos into meaningful classes like indoor, outdoor, mixed, or floorplan, that are commonly present in real-world house listings. These categories are created without manual labeling by aligning image clusters with carefully designed pseudo text prompts. We fuse all features into a unified representation and use an optimized regressor to predict prices. Experiments on a large real-world property dataset demonstrate that our method outperforms traditional methods and existing multimodal baselines. Ablation studies confirm that the novel textual feature extraction and categorized visual embeddings substantially contribute to more accurate and reliable price predictions.

## Kolmogorov-Arnold Networks vs. Multi-Layer Perceptrons: A Rigorous Statistical Benchmark on Clinical Tabular Data

DevFeed: [Kolmogorov-Arnold Networks vs. Multi-Layer Perceptrons: A Rigorous Statistical Benchmark on Clinical Tabular Data](<https://devfeed.tech/articles/kolmogorov-arnold-networks-vs-multi-layer-perceptrons-a-rigorous-statistical-benchmark-on-clinical-tabular-data-114746.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.02856>)

Author: Rafa{\\l} Wodzi\\'nski, Pawe{\\l} Szczerbiak

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [architectures](<https://devfeed.tech/tags/architectures.md>), [bayesian](<https://devfeed.tech/tags/bayesian.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [comparison](<https://devfeed.tech/tags/comparison.md>), [cross-validation](<https://devfeed.tech/tags/cross-validation.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [hyperparameter-optimization](<https://devfeed.tech/tags/hyperparameter-optimization.md>), [kan](<https://devfeed.tech/tags/kan.md>), [latency](<https://devfeed.tech/tags/latency.md>), [medical](<https://devfeed.tech/tags/medical.md>), [metrics](<https://devfeed.tech/tags/metrics.md>), [mlp](<https://devfeed.tech/tags/mlp.md>), [performance](<https://devfeed.tech/tags/performance.md>), [precision](<https://devfeed.tech/tags/precision.md>), [random](<https://devfeed.tech/tags/random.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>)

### AI overview

Across seven clinical tabular datasets, nine KAN variants showed no systematic pooled discrimination advantage over a tuned StandardMLP, though TabKAN and ReLUKAN achieved higher AUPRC on severely imbalanced datasets and could improve minority-class detection. The benchmark used repeated stratified cross-validation with nested Bayesian tuning, assessed AUPRC, Brier Score, and efficiency, and applied Bayesian equivalence testing. KANs had 1.20-1.48x longer per-sample inference latency and up to 12.5x more trainable parameters; StandardMLP remained an efficiency-focused baseline, while KANs may suit tasks prioritizing precision on rare positives.

### Source excerpt

arXiv:2611.02856v1 Announce Type: new Abstract: Kolmogorov-Arnold Networks (KANs) have recently been proposed as a promising alternative to Multi-Layer Perceptrons (MLPs). However, empirical evaluations on structured medical datasets remain scarce, and commonly reported metrics like AUROC can yield overly optimistic performance estimates on imbalanced clinical cohorts. In this study, we present a standardized benchmark comparing eleven architectures - nine KAN variants (FastKAN, WavKAN, ReLUKAN, ChebyKAN, TaylorKAN, GramKAN, JacobiKAN, LegendreKAN, and TabKAN), a tuned StandardMLP, and a Random Forest baseline - across seven heterogeneous medical tabular datasets spanning oncology, cardiology, nephrology, neurology, endocrinology, and obstetrics. Our experimental protocol employs a 3x5 Repeated Stratified K-Fold cross-validation scheme with nested Bayesian hyperparameter optimization and evaluates discrimination via Area Under the Precision-Recall Curve (AUPRC), calibration through the Brier Score, and computational efficiency. Formal equivalence testing is conducted using the Bayesian Correlated t-Test with a Region of Practical Equivalence (ROPE) at +/- 0.01. The results indicate that on datasets with severe class imbalance, dedicated KAN architectures (particularly TabKAN and ReLUKAN) can improve minority-class detection and achieve higher AUPRC than standard MLPs. Nevertheless, across pooled datasets, KAN variants demonstrate comparable overall predictive discrimination rather than a systematic advantage, while requiring 1.20x to 1.48x longer inference latency per sample and up to 12.5x more trainable parameters. Consequently, StandardMLP remains a strong baseline on the efficiency Pareto frontier, while KAN variants provide a viable alternative when precision on rare positive cases is the primary clinical objective. The code for reproducing the results is available at https://github.com/rafalwodzinski/kan_mlp_comparison.

## Why Tabular Foundation Models Excel on Relational Data They Have Never Seen

DevFeed: [Why Tabular Foundation Models Excel on Relational Data They Have Never Seen](<https://devfeed.tech/articles/why-tabular-foundation-models-excel-on-relational-data-they-have-never-seen-113657.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.02936>)

Author: Linjie Xu, Yiqun Hu, Jixiang He, David Wipf

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [rdbms](<https://devfeed.tech/topics/rdbms.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>)

Tags: [agentic-systems](<https://devfeed.tech/tags/agentic-systems.md>), [column](<https://devfeed.tech/tags/column.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [database](<https://devfeed.tech/tags/database.md>), [distributions](<https://devfeed.tech/tags/distributions.md>), [effects](<https://devfeed.tech/tags/effects.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [future](<https://devfeed.tech/tags/future.md>), [interpretability](<https://devfeed.tech/tags/interpretability.md>), [learning](<https://devfeed.tech/tags/learning.md>), [models](<https://devfeed.tech/tags/models.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [properties](<https://devfeed.tech/tags/properties.md>), [table](<https://devfeed.tech/tags/table.md>), [target](<https://devfeed.tech/tags/target.md>), [zero](<https://devfeed.tech/tags/zero.md>)

### AI overview

The paper reports that single-table tabular foundation models can predict target values in relational databases despite having no relational-data pretraining, when paired with parameter-free relational encoders. It confirms this performance across a broader set of experiments and proposes complementary explanations: the encoders turn relational effects into fixed-length embeddings that loosely resemble the distributions the models were designed for, allowing some single-table properties, including convergent posterior predictive distributions, to carry over. The authors suggest this analysis may help guide interpretable relational foundation models and their integration into agentic systems.

### Source excerpt

arXiv:2611.02936v1 Announce Type: cross Abstract: Given a relational database (RDB) storing heterogeneous tabular information, how can we predict missing (or future) values in some target column of interest? As candidate targets are numerous in enterprise settings, it is preferable to avoid learning a new model from scratch each time there is a new prediction task. Frozen foundation models (FMs) based on in-context learning (ICL) and RDB-specific pre-training provide a viable solution, but the ideal design still remains largely unresolved. Perhaps unexpectedly, preliminary evidence has recently shown that existing single-table FMs (aka TFMs), when combined with suitable parameter-free relational encoders, can be quite competitive in practice despite zero exposure to RDB pre-training or relational data of any kind. In this work we first empirically confirm this phenomenon with a more diverse array of experiments. We then turn to our main contribution, namely, surfacing multiple complementary explanations for why TFMs behave so effectively. These explanations ultimately hinge on innate synergies whereby parameter-free encoders map relational effects to fixed-length embeddings that loosely mimic distributions for which TFMs were originally designed; in so doing, desirable single-table properties transfer to relational settings (e.g., convergent posterior predictive distributions). Collectively, our analysis forges new pathways for improving RDB FMs, particularly when we prioritize interpretability and inevitable integration within agentic systems.

## An Agentic AI Workflow for Liver Lesion Segmentation and Structured Radiology Reporting

DevFeed: [An Agentic AI Workflow for Liver Lesion Segmentation and Structured Radiology Reporting](<https://devfeed.tech/articles/an-agentic-ai-workflow-for-liver-lesion-segmentation-and-structured-radiology-reporting-111469.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.34599>)

Author: Fatma Ahmed Mansour, Rahatara Ferdousi, Manar Fathy Elshenawy

Published: 2026-11-02T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Medical imaging](<https://devfeed.tech/topics/medical-imaging.md>), [Agentic AI](<https://devfeed.tech/topics/what-is-agentic-ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentic-ai-framework](<https://devfeed.tech/tags/agentic-ai-framework.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-framework](<https://devfeed.tech/tags/ai-framework.md>), [ai-workflow](<https://devfeed.tech/tags/ai-workflow.md>), [automated](<https://devfeed.tech/tags/automated.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [counts](<https://devfeed.tech/tags/counts.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dimensions](<https://devfeed.tech/tags/dimensions.md>), [framework](<https://devfeed.tech/tags/framework.md>), [generation](<https://devfeed.tech/tags/generation.md>), [image](<https://devfeed.tech/tags/image.md>), [imaging](<https://devfeed.tech/tags/imaging.md>), [language](<https://devfeed.tech/tags/language.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [large](<https://devfeed.tech/tags/large.md>), [reporting](<https://devfeed.tech/tags/reporting.md>), [structured](<https://devfeed.tech/tags/structured.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

### AI overview

The framework links liver and lesion segmentation to traceable measurements and structured radiology reporting. TotalSegmentator-based nnU-Net models delineate the liver, Couinaud segments, and lesions; computational tools extract volumes, dimensions, counts, and locations; and an LLM places those measurements into a predefined schema with numerical consistency checks against source measurements. On 51 LiTS cases, segmentation achieved a liver Dice score of 0.947, lesion sensitivity of 75.6%, and positive predictive value of 68.9%. Exploratory screening of 25,494 MERLIN reports found liver-related abnormalities in 10,208 (40.0%).

### Source excerpt

arXiv:2610.34599v2 Announce Type: replace-cross Abstract: Despite advances in automated liver lesion segmentation, a gap remains between image-derived predictions and clinically interpretable radiology reports. Bridging this gap requires quantitative measurements that are both reproducible and traceable. We present an evidence-grounded agentic AI framework that connects liver and lesion segmentation with deterministic quantification and structured reporting. TotalSegmentator-based nnU-Net models delineate the liver, Couinaud segments, and lesions, while computational tools extract volumes, dimensions, counts, and anatomical locations. A large language model (LLM) organizes these measurements into a predefined reporting schema, with numerical consistency checks against source measurements. On 51 LiTS cases, the segmentation stage achieved a liver Dice score of 0.947, lesion sensitivity of 75.6%, and positive predictive value of 68.9%. Exploratory screening of 25,494 MERLIN reports identified 10,208 (40.0%) with liver-related abnormalities. By connecting quantitative imaging evidence with agentic report generation, this work takes a step toward human-centered medical AI that prioritizes traceability and clinical relevance.

## Beyond IID: How General Are Tabular Foundation Models, Really?

DevFeed: [Beyond IID: How General Are Tabular Foundation Models, Really?](<https://devfeed.tech/articles/beyond-iid-how-general-are-tabular-foundation-models-really-111419.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.30410>)

Author: Lennart Purucker, Andrej Tschalzev, Nick Erickson, Gioia Blayer, David Holzm\\"uller, Alan Arazi, Alexander Pfefferle, Mustafa Tajjar, Ga\\"el Varoquaux, Frank Hutter

Published: 2026-11-02T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Model Development](<https://devfeed.tech/topics/model-development.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [framework](<https://devfeed.tech/tags/framework.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>)

### AI overview

BeyondArena evaluates tabular models across IID, temporal, and grouped tasks, varied dataset sizes and feature dimensions, and data with text or high-cardinality features. The authors introduce DataFoundry, a Python framework and metadata schema for curating datasets, to address fragmented benchmark software and protocols. In tests of 11 models on 142 datasets, open-source tabular foundation models were competitive or excelled on datasets up to 100,000 rows; traditional tree-based and deep learning models remained dominant on grouped or large data. The abstract presents this as evidence that standard IID benchmarks miss harder settings where current foundation models are less successful.

### Source excerpt

arXiv:2606.30410v2 Announce Type: replace Abstract: Foundation models for predictive machine learning on tabular data have recently gained significant traction in academia and industry. Research communities across disciplines are increasingly evaluating tabular foundation models on diverse datasets and tasks. However, these task- and discipline-specific evaluations remain largely inaccessible to model researchers because benchmark software and evaluation protocols are fragmented. As a result, model researchers rely on standard benchmarks, which are mostly defined for tasks where tabular foundation models already excel. The most challenging scenarios are excluded, limiting meaningful progress in the field by focusing on marginal improvements on IID data rather than on broader, more demanding challenges. To overcome this, we introduce BeyondArena, the first unified holistic benchmark for tabular data that supports diverse task types (IID, temporal, grouped), across sample size and feature dimensionality scales, with diverse feature types (with text, with high cardinality) from a broad range of disciplines. To enable unified benchmarking beyond standard benchmarks, we introduce DataFoundry, a Python framework and metadata schema for curating tabular datasets for predictive machine learning. Our results across 11 models and 142 curated datasets show that current open-source tabular foundation models are competitive or excel on datasets up to 100k rows, while traditional tree-based and deep learning models still dominate on grouped or large data. BeyondArena guides model research for the most demanding challenges in tabular data, enabling progress towards truly foundational tabular models.

## Adaptive Allele-Haplotype Incidence Indexing for Pangenome Queries: Representation, Reconstruction, and Benchmarking

DevFeed: [Adaptive Allele-Haplotype Incidence Indexing for Pangenome Queries: Representation, Reconstruction, and Benchmarking](<https://devfeed.tech/articles/adaptive-allele-haplotype-incidence-indexing-for-pangenome-queries-representation-reconstruction-and-benchmarking-108334.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2512.21320>)

Author: Roberto Garrone

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [Compression](<https://devfeed.tech/topics/compression.md>), [Encoding](<https://devfeed.tech/topics/encoding.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Phoronix Test Suite](<https://devfeed.tech/topics/phoronix-test-suite.md>), [Graph Mining](<https://devfeed.tech/topics/graph-mining.md>)

Tags: [22](<https://devfeed.tech/tags/22.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [archive](<https://devfeed.tech/tags/archive.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [assertions](<https://devfeed.tech/tags/assertions.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [but](<https://devfeed.tech/tags/but.md>), [canonical](<https://devfeed.tech/tags/canonical.md>), [carrier](<https://devfeed.tech/tags/carrier.md>), [catalog](<https://devfeed.tech/tags/catalog.md>), [comparison](<https://devfeed.tech/tags/comparison.md>), [compression](<https://devfeed.tech/tags/compression.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-ds](<https://devfeed.tech/tags/cs-ds.md>), [custom](<https://devfeed.tech/tags/custom.md>), [edge](<https://devfeed.tech/tags/edge.md>), [indexing](<https://devfeed.tech/tags/indexing.md>), [q-bio-gn](<https://devfeed.tech/tags/q-bio-gn.md>), [queries](<https://devfeed.tech/tags/queries.md>)

### AI overview

The study tests two representations for phased-variant queries: H1 stores adaptive allele-haplotype incidence rows, while H2 encodes ordered edge paths from a fixed, nonoverlapping variant catalog. In a chromosome 1 interval, H1's adaptive files used 0.872 MiB (0.207 MiB after whole-file Zstandard compression), compared with 1.126 MiB for a custom Roaring-like layout and 1.190 MiB for a bitmap-payload baseline. For a separate 22,727-site SNV projection, context-Huffman coding reduced H2 edge sequences from 75,130,882 to 5,156,418 bytes. Notebook carrier-equivalence and graph-adjacency checks passed, and reconstruction is established conditionally for abstract H2 walks. The results show restricted implementation feasibility; cross-callset phasing, nucleotide-level structural-variant reconstruction, and comparison with production indexes remain unresolved, so they do not establish superior end-to-end pangenome compression.

### Source excerpt

arXiv:2512.21320v3 Announce Type: replace-cross Abstract: Population-scale analysis of phased genetic variants requires both allele-carrier retrieval and ordered haplotype context. We examine H1, an adaptive allele-haplotype incidence-row index, and H2, a canonical edge-path representation derived from a fixed, nonoverlapping variant catalog. We prove conditional reconstruction of abstract H2 walks and distinguish this property from unrestricted equivalence to genomic variation graphs. A completed local notebook experiment used a chromosome 1 interval (20-22 Mb), 200 diploid individuals, 24,921 SNV/indel records, and 45 SVTYPE-annotated records. For 24,966 extracted nonreference rows across 400 haplotypes, the serialized H1 incidence-row files occupied 0.872 MiB with a simple adaptive layout and 1.126 MiB with a custom Roaring-like layout, versus 1.190 MiB for a bitmap-payload baseline; a whole-file Zstandard archive of the adaptive layout occupied 0.207 MiB. For a separately constructed nonoverlapping SNV projection containing 22,727 sites, indexed H2 edge sequences occupied 75,130,882 bytes in a raw format and 5,156,418 bytes with context-Huffman encoding (14.57-fold reduction). Carrier-equivalence and graph-adjacency assertions passed in the notebook, but full crosscallset phasing, nucleotide-level SV reconstruction, and comparison with production indexes remain unresolved. The measurements establish a restricted implementation feasibility result, not superior end-to-end pangenome compression.

## Octopus: Agentic Multi-Table Discovery over Tabular Data Lakes

DevFeed: [Octopus: Agentic Multi-Table Discovery over Tabular Data Lakes](<https://devfeed.tech/articles/octopus-agentic-multi-table-discovery-over-tabular-data-lakes-108335.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2601.02304>)

Author: Wen-Zhi Li, Sainyam Galhotra

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [Structured-data](<https://devfeed.tech/topics/structured-data.md>), [Agentic AI](<https://devfeed.tech/topics/what-is-agentic-ai.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [schema-evolution](<https://devfeed.tech/topics/schema-evolution.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [api](<https://devfeed.tech/tags/api.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [code](<https://devfeed.tech/tags/code.md>), [column](<https://devfeed.tech/tags/column.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [data-discovery](<https://devfeed.tech/tags/data-discovery.md>), [data-lakes](<https://devfeed.tech/tags/data-lakes.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [discovery](<https://devfeed.tech/tags/discovery.md>), [execution](<https://devfeed.tech/tags/execution.md>), [experimental](<https://devfeed.tech/tags/experimental.md>), [framework](<https://devfeed.tech/tags/framework.md>), [octopus](<https://devfeed.tech/tags/octopus.md>), [table](<https://devfeed.tech/tags/table.md>), [tabular-data](<https://devfeed.tech/tags/tabular-data.md>)

### AI overview

Octopus is a lightweight agentic framework for finding relevant tables when a natural-language question may require information from multiple tables. It combines lightweight offline semantic guidance with query-time access to raw tabular data; an LLM agent orchestrates semantic column search, structured tabular operations, and general-purpose execution, using intermediate observations to locate and verify tables. On a new benchmark spanning four independent-discovery and two join-based datasets, Octopus achieved the highest macro-F1 on all six. The abstract reports lower offline preprocessing overhead than Pneuma and lower online latency and API cost than Claude Code. With the same NL2SQL agent, it also produced substantially higher downstream question-answering F1 than Claude Code. No scores or evaluation details are provided in the supplied excerpt.

### Source excerpt

arXiv:2601.02304v2 Announce Type: replace Abstract: Tabular data discovery aims to identify the tables relevant to a natural-language question from a large collection of heterogeneous data. Existing systems typically assume that a question is answered by a single table and rely on retrieval pipelines constructed through substantial offline preprocessing. These assumptions become restrictive when relevant information is distributed across multiple tables. In this paper, we present Octopus, a lightweight agentic framework for multi-table data discovery. Octopus follows a thin-harness design that combines lightweight offline semantic guidance with adaptive access to raw tabular data. At query time, an LLM agent dynamically orchestrates semantic column search, structured tabular operations, and general-purpose execution to locate and verify relevant tables based on intermediate observations. We further introduce a benchmark that explicitly evaluates multi-table data discovery, covering both settings where multiple tables independently answer a question and settings where multiple tables must be joined. Experimental results across four independent-discovery datasets and two join-based datasets show that Octopus achieves the highest macro-F1 on all six datasets. It also substantially reduces offline preprocessing overhead relative to Pneuma and online latency and API cost relative to Claude Code. When coupled with the same NL2SQL agent, Octopus yields substantially higher downstream question-answering F1 than Claude Code, showing that improved table discovery provides more complete evidence for downstream analytics. Code is available at https://github.com/wenzhilics/octopus.

[Next page](<https://devfeed.tech/topics/structured-data.md?cursor=WyIyMDI2LTEwLTMwVDA0OjAwOjAwKzAwOjAwIiwgIjQ4ZWFmY2M2LTJlZDUtNGE5My1hNWJjLWFkNzZjMGIyMzgyYyJd>)