# TensorFlow Decision Forests

TensorFlow Decision Forests (TF-DF) is a TensorFlow library for training, running, and interpreting decision forest models, supporting classification, regression, and ranking.

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

## Lost in the Funnel: Selection-Induced Miscalibration in Ranking and Decision Systems

DevFeed: [Lost in the Funnel: Selection-Induced Miscalibration in Ranking and Decision Systems](<https://devfeed.tech/articles/lost-in-the-funnel-selection-induced-miscalibration-in-ranking-and-decision-systems-141790.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22595>)

Author: Aditya Somani, Rahul Ragesh, Vikas Joshi, Amit Singh

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

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>), [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>)

Tags: [auction](<https://devfeed.tech/tags/auction.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [business](<https://devfeed.tech/tags/business.md>), [case](<https://devfeed.tech/tags/case.md>), [component](<https://devfeed.tech/tags/component.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [decision](<https://devfeed.tech/tags/decision.md>), [diagnosis](<https://devfeed.tech/tags/diagnosis.md>), [diagnostic](<https://devfeed.tech/tags/diagnostic.md>), [effects](<https://devfeed.tech/tags/effects.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [error](<https://devfeed.tech/tags/error.md>), [framework](<https://devfeed.tech/tags/framework.md>), [funnel](<https://devfeed.tech/tags/funnel.md>), [industry](<https://devfeed.tech/tags/industry.md>), [layer](<https://devfeed.tech/tags/layer.md>), [live](<https://devfeed.tech/tags/live.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [selection](<https://devfeed.tech/tags/selection.md>), [systems](<https://devfeed.tech/tags/systems.md>)

### AI overview

Scores calibrated at a model boundary can become miscalibrated after filtering, reranking, business rules, or auctions, even when selection rather than model error causes the deterioration. The paper diagnoses this with boundary-wise residual identities, measuring calibration against rubric-judged relevance rather than presentation-dependent metrics such as CTR. It proposes Funnel-Aware Recalibration (FAR), a post-hoc method that accounts for the selection policy at the decision boundary. In a live two-region production A/B test, FAR reduced decision-time expected calibration error by 60-80% and improved auction efficiency. Adapted public datasets provide further evidence for the proposed mechanism; the authors say they plan to release them and their generation pipeline.

### Source excerpt

arXiv:2611.22595v1 Announce Type: new Abstract: Modern ranking and decision systems combine components with divergent objectives, and scores calibrated at the model boundary are used as calibrated probabilities throughout the funnel, across filtering, reranking, business-rule enforcement, and auctions. Yet calibration is typically validated only where the score is produced, so its breakdown where the score is used is routinely overlooked. We ask when calibration survives this selection, how inherited error differs from selection effects, and what can fail for a perfectly calibrated score at a component boundary. We study these questions through a boundary-wise diagnostic framework of residual identities. We target rubric-judged relevance rather than presentation-dependent metrics such as CTR, so calibration is measurable at every funnel layer of interest. We first present an industry-wide case study that applies our diagnosis to settings where calibration deteriorates at decision time, tracing the degradation to selection rather than model error. This diagnosis motivates Funnel-Aware Recalibration (FAR): post-hoc, selection-policy-aware recalibration at the boundary where scores drive decisions. In a live two-region production A/B test, FAR cuts decision-time expected calibration error (ECE) by 60-80% and improves auction efficiency. We further adapt two public datasets to simulate industry settings, providing additional evidence for the mechanism. We plan to release the adapted datasets and their generation pipeline to support further research. Finally, we distill these findings into practical operating principles for calibrating multi-stage scoring systems.

## Random Cut Forests, Games and Diffusion over Approximately Nested Anisotropic Local Geometry

DevFeed: [Random Cut Forests, Games and Diffusion over Approximately Nested Anisotropic Local Geometry](<https://devfeed.tech/articles/random-cut-forests-games-and-diffusion-over-approximately-nested-anisotropic-local-geometry-138885.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.19548>)

Author: Sudipto Guha

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

Content type: article

Language: en

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

Topics: [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>), [stream-processing](<https://devfeed.tech/topics/stream-processing.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [added](<https://devfeed.tech/tags/added.md>), [affinity](<https://devfeed.tech/tags/affinity.md>), [align](<https://devfeed.tech/tags/align.md>), [also](<https://devfeed.tech/tags/also.md>), [ambient](<https://devfeed.tech/tags/ambient.md>), [approach](<https://devfeed.tech/tags/approach.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [axis](<https://devfeed.tech/tags/axis.md>), [based](<https://devfeed.tech/tags/based.md>), [basis](<https://devfeed.tech/tags/basis.md>), [but](<https://devfeed.tech/tags/but.md>), [calculus](<https://devfeed.tech/tags/calculus.md>), [change](<https://devfeed.tech/tags/change.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [games](<https://devfeed.tech/tags/games.md>), [local](<https://devfeed.tech/tags/local.md>), [nested](<https://devfeed.tech/tags/nested.md>), [over](<https://devfeed.tech/tags/over.md>), [random](<https://devfeed.tech/tags/random.md>)

### AI overview

The paper uses the evolving geometry of random cut forests as a shared energy framework for streaming prediction, explanation, games, and diffusion. Existing scores are interpreted as random-walk escape probabilities, and the resulting affinity is represented exactly by a sparse electrical network on each tree, maintained along one path per update. Different clamp choices yield harmonic predictions with Gaussian uncertainty, normalized cuts with costs attributed to oriented faces, or coalition games that measure field response. With noise, the energy also defines diffusion dynamics whose relaxation rates come from the boxes. The authors state that localization of axis-parallel cuts is controlled by affine rather than ambient dimension; across a forest, boxes are only approximately nested and are aligned using common reference observations. A noisy rotating-star example illustrates the measurements.

### Source excerpt

arXiv:2611.19548v1 Announce Type: new Abstract: Just as the data in a dynamic stream changes, questions asked of the stream also change in intent: what to hold constant, which counterfactuals to consider, and so forth. Energy-based learning has been celebrated for its uniform approach to losses and estimation, but interpretability remains challenging. Tree-based methods, on the other hand, offer high interpretability but often rely on bespoke, disjointed estimators. Neither paradigm readily supports changing questions on streaming data. In this paper we take a step towards combining energy, interpretability, and streaming. Random cut forests retain an evolving geometry of the data. We study this geometry as the basis for diffusion dynamics, prediction, explanation, and games. We show that several existing scores are escape probabilities of random walks on the trees. The induced affinity is realized exactly by a sparse electrical network on each tree, giving a single energy calculus, a Hamiltonian, maintained along one path per update. This connects an estimate, its supporting evidence, and the response to intervention through the same boxes, gaps, masses, and paths. Different questions correspond to different clamp decisions on the energy. Fixing observed values gives harmonic prediction with Gaussian uncertainty. Fixing a partition gives normalized cuts whose cost is attributed exactly to oriented faces. Releasing clamps gives coalition games, in which collusive displacement measures the field response. With noise, the same energy gives diffusion dynamics whose relaxation rates are read from the boxes. Axis-parallel cuts localize at a rate set by affine rather than ambient dimension, and this control passes to the averaged energy. Across a forest the boxes are only approximately nested; common reference observations align the trees. A rotating star with added noise illustrates the measurements.

## Explaining Cross-Label Decision Structure in Multi-Output Tree Ensembles

DevFeed: [Explaining Cross-Label Decision Structure in Multi-Output Tree Ensembles](<https://devfeed.tech/articles/explaining-cross-label-decision-structure-in-multi-output-tree-ensembles-137643.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.20996>)

Author: Sylvio Barbon Junior, Hiago F. F. Oliveira, Rafael Figueira Gon\\c{c}alves, Jessica Fernandes Lopes, Ricardo Cerri

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

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: [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>), [model architecture](<https://devfeed.tech/topics/model-architecture.md>), [Graphs](<https://devfeed.tech/topics/graphs.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [attribution](<https://devfeed.tech/tags/attribution.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [condition](<https://devfeed.tech/tags/condition.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [decision](<https://devfeed.tech/tags/decision.md>), [explain](<https://devfeed.tech/tags/explain.md>), [flow](<https://devfeed.tech/tags/flow.md>), [graph](<https://devfeed.tech/tags/graph.md>), [information](<https://devfeed.tech/tags/information.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [predicate](<https://devfeed.tech/tags/predicate.md>), [probability](<https://devfeed.tech/tags/probability.md>), [rankings](<https://devfeed.tech/tags/rankings.md>), [share](<https://devfeed.tech/tags/share.md>), [show](<https://devfeed.tech/tags/show.md>), [signal](<https://devfeed.tech/tags/signal.md>), [structure](<https://devfeed.tech/tags/structure.md>)

### AI overview

The proposed multi-sink Decision Predicate Graphs (ms-DPGs) and Shared Predicate Flow (SPF) explain how multi-output tree ensembles reuse decision conditions across labels. SPF measures how much of a label's decision mass passes through conditions active for both labels; a contracted graph view uses label-specific lifts to show whether shared conditions raise both labels' probabilities or favor one over the other. Controlled experiments indicate SPF captures shared decision conditions, not just feature overlap. Across nine benchmarks, native multi-output forests shared more decision structure than Binary Relevance on eight, including five of six cases at predictive parity; per-label TreeSHAP rankings showed architectural changes but not the degree of sharing. A key limitation: SPF describes the fitted model, and rose when feature-label signal was removed on eight benchmarks, so high SPF does not establish shared predictive information.

### Source excerpt

arXiv:2611.20996v1 Announce Type: cross Abstract: A multi-output tree ensemble predicts several binary labels through decision paths that it largely shares across those labels, yet the attribution methods commonly used to explain such models treat each output in isolation. We propose multi-sink Decision Predicate Graphs (ms-DPGs), which extend decision predicate graphs with one sink per label: each leaf contributes to the sink of every label its in-bag positive count for that label. On this representation we define Shared Predicate Flow (SPF), the share of a label's decision mass that passes through conditions active for both labels of a pair, and a contracted view of the graph in which label-specific lifts indicate whether a shared condition raises the probability of both labels or of one at the expense of the other. Controlled experiments show that SPF captures shared decision conditions rather than mere feature overlap. On nine multi-label benchmarks, native multi-output forests share substantially more decision structure than Binary Relevance on eight, including five of the six on which the two are at predictive parity, even at equal hyperparameters, whereas per-label TreeSHAP rankings register the change of architecture without showing whether more is shared. Because a native forest routes every label through the same splits, SPF describes the fitted model rather than the data, and on eight of the nine benchmarks it rises when the feature-label signal is removed; a high SPF is therefore not evidence of shared predictive information. ms-DPG offers a global, predicate-level account of how a multi-output ensemble reuses decision conditions across its labels.

## Global-Local and Window-Concatenated Functional Random Forests: Early and Late Fusion of Window-Specific Functional Principal Components

DevFeed: [Global-Local and Window-Concatenated Functional Random Forests: Early and Late Fusion of Window-Specific Functional Principal Components](<https://devfeed.tech/articles/global-local-and-window-concatenated-functional-random-forests-early-and-late-fusion-of-window-specific-functional-principal-components-133042.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.15816>)

Author: Fabrizio Maturo, Annamaria Porreca

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

Content type: article

Language: en

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

Topics: [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>), [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [functional](<https://devfeed.tech/tags/functional.md>), [fusion](<https://devfeed.tech/tags/fusion.md>), [global](<https://devfeed.tech/tags/global.md>), [local](<https://devfeed.tech/tags/local.md>), [machine](<https://devfeed.tech/tags/machine.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [random](<https://devfeed.tech/tags/random.md>), [stat-me](<https://devfeed.tech/tags/stat-me.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [statistics](<https://devfeed.tech/tags/statistics.md>), [window](<https://devfeed.tech/tags/window.md>)

### AI overview

The paper proposes two functional random forests that use window-specific functional principal components to capture features that a whole-domain decomposition may dilute. GL-FRF uses late fusion: it combines a tuned full-curve forest and window-specific forests with penalised convex weights. The full-curve forest remains a feasible solution, and the weights indicate which regions contribute to predictions. WC-FRF uses early fusion, training one tuned forest on concatenated window-specific scores. Across 20 UCR datasets, GL-FRF improved on a same-tree-count tuned full-curve forest on 16, with Brier score reductions up to 0.21 that were significant after Holm adjustment. WC-FRF was more accurate than GL-FRF on 13 datasets but failed when prediction required the full curve. Across 32 simulated and real settings, GL-FRF had the smallest mean and maximum excess over the best method; no method was best on more than six datasets.

### Source excerpt

arXiv:2611.15816v1 Announce Type: cross Abstract: Functional random forests built on a single functional principal component decomposition of the whole domain may dilute features confined to short intervals, whose location is rarely known in advance. We propose two random forests based on window-specific functional principal components. The global-local functional random forest (GL-FRF) is a late-fusion method that stacks a tuned full-curve forest and one forest per window with penalised convex weights; it contains the full-curve forest as a feasible solution, and its weights indicate which parts of the domain drive the prediction. The window-concatenated functional random forest (WC-FRF) is an early-fusion method that trains a single tuned forest on the concatenated window-specific scores. On twenty UCR datasets, GL-FRF improved on a tuned full-curve forest with the same number of trees on 16, with Brier score reductions of up to 0.21, significantly after Holm adjustment. No method was best on more than six datasets. WC-FRF was more accurate than GL-FRF on 13 datasets but failed when the signal required the full curve, whereas GL-FRF had the smallest mean and maximum excess over the best method across all 32 simulated and real settings, which makes it the safer choice when the location of the signal is unknown.

## Vectorized Dynamic Histograms for Sparse Oblique Forests

DevFeed: [Vectorized Dynamic Histograms for Sparse Oblique Forests](<https://devfeed.tech/articles/vectorized-dynamic-histograms-for-sparse-oblique-forests-126682.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2603.00326>)

Author: Ariel Lubonja, Jungsang Yoon, Haoyin Xu, Yue Wan, Yilin Xu, Richard Stotz, Mathieu Guillame-Bert, Joshua T. Vogelstein, Randal Burns

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

Content type: article

Language: en

Sources: [arXiv: Distributed Computing](<https://devfeed.tech/sources/arxiv-distributed-computing.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>)

Tags: [avx](<https://devfeed.tech/tags/avx.md>), [avx2](<https://devfeed.tech/tags/avx2.md>), [cs-dc](<https://devfeed.tech/tags/cs-dc.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-pf](<https://devfeed.tech/tags/cs-pf.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [gradient-boosted-trees](<https://devfeed.tech/tags/gradient-boosted-trees.md>), [linear-combination](<https://devfeed.tech/tags/linear-combination.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [novel](<https://devfeed.tech/tags/novel.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [optimizations](<https://devfeed.tech/tags/optimizations.md>), [rf](<https://devfeed.tech/tags/rf.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [scanning](<https://devfeed.tech/tags/scanning.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>), [time](<https://devfeed.tech/tags/time.md>), [top](<https://devfeed.tech/tags/top.md>), [training](<https://devfeed.tech/tags/training.md>), [trees](<https://devfeed.tech/tags/trees.md>)

### AI overview

The paper reports substantial training speedups for sparse oblique forests in Yggdrasil Decision Forests (YDF), where runtime-sampled feature projections prevent presorting. Fixing inefficiencies in the released implementation yields 2-5x speedup. Hierarchical AVX2 and AVX-512 histogram filling improves typical-depth SPO-GBT by up to 1.6x and SPO-RF by up to 2x; runtime-dynamic histograms, which bridge vectorized histograms and exact splits, improve deeper SPO trees by up to 1.6x with no detectable accuracy effect. The resulting SPO-RF training time is reported as on par with YDF's axis-aligned RF. Evaluation covered 8 natural and 11 synthetic datasets, up to 10.5 million rows and 1.6 million features, from depth 6 to purity; the implementation is open source.

### Source excerpt

arXiv:2603.00326v3 Announce Type: replace-cross Abstract: Sparse oblique (SPO), part of the top-ranked configuration of Google's Yggdrasil Decision Forests (YDF), improve the accuracy while maintaining interpretability of Random Forests (RF) and Gradient Boosted Trees (GBT) by scanning a sparse linear combination of features instead of a single feature. Because projections are sampled at runtime, training is significantly slower, as pre-run optimizations such as presorting cannot be used. We overhaul SPO training in YDF, first by fixing inefficiencies in the released version, achieving 2-5x speedup, then by devising two novel methods, one aimed at GBTs and the other at RFs: (1) Hierarchical AVX2 and AVX-512-vectorized histogram filling speeds up typical GBT-depth trees by up to 1.6x in SPO-GBT and 2x for SPO-RF, and (2) runtime-dynamic histograms between vectorized histograms and exact splits, speeds up deeper, RF depth-typical trees (depths 16 to purity) by up to 1.6x with no detectable effect on accuracy. Our optimizations bring SPO-RF training time on par with YDF's axis-aligned RF. We extensively evaluate on 8 natural and 11 synthetic datasets of up to 10.5 million rows and 1.6 million features, from depth 6 to purity, and open-source the implementation.

## When Does Combining Software-Effort Estimates Help? Evidence from Four Public Project Datasets

DevFeed: [When Does Combining Software-Effort Estimates Help? Evidence from Four Public Project Datasets](<https://devfeed.tech/articles/when-does-combining-software-effort-estimates-help-evidence-from-four-public-project-datasets-120798.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.06948>)

Author: Azzeddine Ihsine, Sara Ihsine

Published: 2026-11-06T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Software](<https://devfeed.tech/topics/software.md>), [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>)

Tags: [combined](<https://devfeed.tech/tags/combined.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [covariance-shrinkage](<https://devfeed.tech/tags/covariance-shrinkage.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [effort](<https://devfeed.tech/tags/effort.md>), [error](<https://devfeed.tech/tags/error.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [partitions](<https://devfeed.tech/tags/partitions.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [project](<https://devfeed.tech/tags/project.md>), [random](<https://devfeed.tech/tags/random.md>), [regression](<https://devfeed.tech/tags/regression.md>), [rules](<https://devfeed.tech/tags/rules.md>), [software](<https://devfeed.tech/tags/software.md>), [study](<https://devfeed.tech/tags/study.md>), [test](<https://devfeed.tech/tags/test.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

Across three datasets, combining software-effort estimates reduced mean log-MAE by 1.3-8.6%, but the gains were not statistically distinguishable from zero after accounting for overlap between repeated test splits. On Kitchenham, a recorded human estimate dominated: equal weighting with model estimates significantly worsened mean test log-MAE, while a nonnegative simplex combination limited the loss. On a completion-year split of Desharnais, simplex and covariance-shrinkage combinations retained gains of about 5%, while equal weighting worsened slightly and inverse-MSE weighting was near the individual-model baseline. The study's results suggest that pooling can dilute a strong estimator, and benefits among other correlated estimates were small and uncertain.

### Source excerpt

arXiv:2611.06948v1 Announce Type: new Abstract: Combining estimates can reduce individual prediction errors, but an ensemble may also dilute an unusually accurate estimator or repeat errors shared by its members. We study these competing effects on four archival software-effort datasets (Kitchenham: 145 projects, Desharnais: 77, China: 499, Maxwell: 63) with 100 repeated 70/30 training and test partitions per dataset. Three conventional models (parametric regression, analogy, and random forest) are combined with each other and, in Kitchenham, with a recorded human estimate. Out-of-fold calibration errors drive five combination rules: equal weighting, inverse-MSE weighting, a nonnegative simplex combination, signed Bates-Granger weighting, and Bates-Granger with covariance shrinkage. All preprocessing is learned within training folds. Model-error correlations range from 0.54 to 0.97. In Kitchenham, the human estimate dominates: equal weighting raises mean test log-MAE from 0.239 to 0.455, a loss that remains significant under a corrected resampled t-test with Holm adjustment, whereas the simplex rule limits the increase to 0.250. In Desharnais, China, and Maxwell, every rule lowers mean log-MAE (by 1.3 to 8.6%) and BG with shrinkage beats the calibration-selected individual on 75, 83, and 83 of 100 splits, but none of these gains is statistically distinguishable from zero once the overlap between repeated splits is accounted for. On a completion-year split of Desharnais, equal weighting becomes slightly worse and inverse-MSE is near the individual baseline, while simplex and shrinkage retain mean gains of about 5%. Combining correlated effort estimates is therefore a conditional decision: the harm of naive pooling next to a dominant source is clear, while the benefit elsewhere is small and uncertain.

## Pre-processing temporal data made easier with TensorFlow Decision Forests and Temporian

DevFeed: [Pre-processing temporal data made easier with TensorFlow Decision Forests and Temporian](<https://devfeed.tech/articles/pre-processing-temporal-data-made-easier-with-tensorflow-decision-forests-and-temporian-7387.md>)

Original publisher: [Read original article](<https://blog.tensorflow.org/2023/09/forecasting-with-tensorflow-decision-forests-and-temporian.html>)

Author: TensorFlow Blog (noreply@blogger.com)

Published: 2023-09-11T20:14:00Z

Content type: article

Language: en

Sources: [The TensorFlow Blog](<https://devfeed.tech/sources/the-tensorflow-blog.md>)

Topics: [Temporal data](<https://devfeed.tech/topics/temporal-data.md>), [Temporian](<https://devfeed.tech/topics/temporian.md>), [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>), [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [Python](<https://devfeed.tech/topics/python.md>), [dataset](<https://devfeed.tech/topics/dataset.md>)

Tags: [blog](<https://devfeed.tech/tags/blog.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [feature-engineering](<https://devfeed.tech/tags/feature-engineering.md>), [logs](<https://devfeed.tech/tags/logs.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [python](<https://devfeed.tech/tags/python.md>), [sales](<https://devfeed.tech/tags/sales.md>), [temporal-data](<https://devfeed.tech/tags/temporal-data.md>), [temporian](<https://devfeed.tech/tags/temporian.md>), [tensorflow-decision-forests](<https://devfeed.tech/tags/tensorflow-decision-forests.md>), [time-sequences](<https://devfeed.tech/tags/time-sequences.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [transactions](<https://devfeed.tech/tags/transactions.md>), [user-event](<https://devfeed.tech/tags/user-event.md>)

### AI overview

This article demonstrates preprocessing temporal sales data with Temporian and forecasting weekly sales with TensorFlow Decision Forests. It explains event sets, aggregation, moving sums, indexing, and the trade-off between transactional detail and uniformly sampled time series.

### Source excerpt

Posted by Google: Mathieu Guillame-Bert, Richard Stotz, Robert Crowe, Luiz GUStavo Martins (Gus), Ashley Oldacre, Kris Tonthat, Glenn Cameron, and Tryolabs: Ian Spektor, Braulio Rios, Guillermo Etchebarne, Diego Marvid, Lucas Micol, Gonzalo Marín, Alan Descoins, Agustina Pizarro, Lucía Aguilar, Martin Alcala Rubi Temporal data is omnipresent in applied machine learning applications. Data often changes over time or is only available or valuable at a certain point in time. For example, market prices and weather conditions change constantly. Temporal data is also often highly discriminative in decision-making tasks. For example, the rate of change and interval between two consecutive heartbeats provides valuable insights into a person's physical health, and temporal patterns of network logs are used to detect configuration issues and intrusions. Hence, it is essential to incorporate temporal data and temporal information in ML applications. INFO: Temporian is a new open-source Python library for preprocessing and feature engineering temporal data for machine learning applications. It is developed in collaboration between Google and Tryolabs. Check the sister blog post for more details. This blog post demonstrates how to train a forecasting model on transactional data. Specifically, we will show how to forecast the total weekly sales from individual sales records. For the modeling part, we will use TensorFlow Decision Forests as they are well suited to handle temporal data. To feed the transaction data to our model, and to compute temporal specific features, we will use Temporian, a newly released library designed for ingesting and aggregating transactional data from multiple non-synchronized sources. Time series are the most commonly used representation for temporal data. They consist of uniformly sampled values, which can be useful for representing aggregate signals. However, time series are sometimes not sufficient to represent the richness of available data. Instead

## Explore-exploit dilemma in Ranking model

DevFeed: [Explore-exploit dilemma in Ranking model](<https://devfeed.tech/articles/explore-exploit-dilemma-in-ranking-model-28029.md>)

Original publisher: [Read original article](<https://tech.trivago.com/post/2022-11-04-explore-exploit-dilemma-in-ranking-model/>)

Author: Aida Orujova Data Scientist

Published: 2022-11-04T00:00:00Z

Content type: article

Language: en

Sources: [Trivago](<https://devfeed.tech/sources/trivago.md>)

Topics: [recommendation systems](<https://devfeed.tech/topics/recommendation-systems.md>), [information retrieval](<https://devfeed.tech/topics/information-retrieval.md>), [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>), [TensorFlow Decision Forests](<https://devfeed.tech/topics/tensorflow-decision-forests.md>), [Test automation](<https://devfeed.tech/topics/test-automation.md>)

Tags: [algorithm](<https://devfeed.tech/tags/algorithm.md>), [cold-start-problem](<https://devfeed.tech/tags/cold-start-problem.md>), [control](<https://devfeed.tech/tags/control.md>), [cost](<https://devfeed.tech/tags/cost.md>), [data](<https://devfeed.tech/tags/data.md>), [data-analysis](<https://devfeed.tech/tags/data-analysis.md>), [data-mining](<https://devfeed.tech/tags/data-mining.md>), [data-preprocessing](<https://devfeed.tech/tags/data-preprocessing.md>), [data-science](<https://devfeed.tech/tags/data-science.md>), [dilemma](<https://devfeed.tech/tags/dilemma.md>), [distribution](<https://devfeed.tech/tags/distribution.md>), [eda](<https://devfeed.tech/tags/eda.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [exploit](<https://devfeed.tech/tags/exploit.md>), [exploitation](<https://devfeed.tech/tags/exploitation.md>), [exploration](<https://devfeed.tech/tags/exploration.md>), [explore](<https://devfeed.tech/tags/explore.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [model](<https://devfeed.tech/tags/model.md>), [ranking](<https://devfeed.tech/tags/ranking.md>)

### AI overview

trivago tested ways to balance showing historically successful accommodations with exposing users to promising, little-seen inventory. Its baseline used the mean of a beta-binomial performance estimate; a naive variant added a tunable fraction of the estimate's standard deviation, favoring items with fewer impressions. Because this could not distinguish equally exposed items, model-based variants used features and high-impression inventory to predict performance among low-impression candidates. Across two quarters of A/B tests, the accepted mechanism increased exposure to high-quality unexplored inventory without short-term revenue cost, and advertiser clickshares did not shift significantly. Lambda provided a control for exploration extent and its trade-off with topline metrics. Results varied: some models explored negatively or had lower quality than the naive approach, so output behavior needed evaluation against the experiment metrics. The mechanism is in production with monitoring for feature distributions, exploration costs, and quality.

### Source excerpt

Imagine, out of thousands of accommodations that match a user search, you have to select the "best" 25 to show to the user. Which ones would you show- the ones you know perform well or ones that...