# Time Series

Time series is an ordered sequence of measurements of a variable at equally spaced time intervals, used in analysis and forecasting.

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

## LightGBM vs. a Four-Week Average for Forecasting Bike-Share Rentals

DevFeed: [LightGBM vs. a Four-Week Average for Forecasting Bike-Share Rentals](<https://devfeed.tech/articles/lightgbm-vs-a-four-week-average-for-forecasting-bike-share-rentals-146443.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/lightgbm-vs-a-four-week-average-for-forecasting-bike-share-rentals/>)

Author: Nahid Ahmadvand

Published: 2026-11-20T16:30:00Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Feature Engineering](<https://devfeed.tech/topics/feature-engineering.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [scikit-learn](<https://devfeed.tech/topics/scikit-learn.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [bike-share](<https://devfeed.tech/tags/bike-share.md>), [editor-s-picks](<https://devfeed.tech/tags/editor-s-picks.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [lightgbm](<https://devfeed.tech/tags/lightgbm.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [python](<https://devfeed.tech/tags/python.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>)

### AI overview

In a 48-hour-ahead Capital Bikeshare forecasting experiment, a four-week seasonal average proved difficult to beat during development: LightGBM's MAE was 49.025 versus 49.856, and paired tests did not establish a difference. On 30 later November-December forecast paths, the selected LightGBM pipeline with lagged weather had 19.6% lower MAE than the baseline, though the author did not test that difference for statistical significance and cannot identify what caused the change. The workflow reconstructed 159 missing development hours as zeros after hourly sums matched daily totals, while distinguishing recorded rentals from actual demand; it excluded nonexistent daylight-saving hours from scoring and constrained features to information available at midnight. Future observed weather sharply reduced error but was unavailable at forecast time. LightGBM relied heavily on weekly summaries, and it struggled during Hurricane Sandy. Empirical prediction intervals came close to target coverage during development but fell short in the later period. The experiment did not compare foundation models.

### Source excerpt

A 48-hour Capital Bikeshare experiment pits XGBoost, Ridge, SARIMA, and Prophet against simple baselines, with care for missing hours and leakage. The post LightGBM vs. a Four-Week Average for Forecasting Bike-Share Rentals appeared first on Towards Data Science.

## Channel-Dependent State Space Model for Multivariate Time Series Forecasting

DevFeed: [Channel-Dependent State Space Model for Multivariate Time Series Forecasting](<https://devfeed.tech/articles/channel-dependent-state-space-model-for-multivariate-time-series-forecasting-144942.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.36453>)

Author: Yu-Cheng Wu, Fan-Keng Sun, Li-Chun Lu, Duane S. Boning

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

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Mamba](<https://devfeed.tech/topics/mamba.md>)

Tags: [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [scalability](<https://devfeed.tech/tags/scalability.md>), [state-space-model](<https://devfeed.tech/tags/state-space-model.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>), [variables](<https://devfeed.tech/tags/variables.md>)

### AI overview

Chameleon combines selective state space models with the Kalman filter's measurement-update formulation to model fine-grained, data-dependent interactions across variables, while retaining an SSM backbone for temporal dynamics and scaling linearly with variable count. It uses GatedDeltaNet and stochastic perturbation of reversible instance normalization to support generalization. On strongly dependent ODE and PEMS datasets, it reports the best MSE and MAE across all settings; across 28 standard benchmark settings, it beats each baseline on MSE in at least 27 and on MAE in at least 22. The abstract reports competitive efficiency and favorable scalability on Traffic and ETT. These are reported results from the supplied abstract.

### Source excerpt

arXiv:2609.36453v3 Announce Type: replace Abstract: Multivariate time series forecasting (MTSF) is critical across many real-world domains. Existing deep learning approaches fall into two paradigms with distinct limitations: channel-independent (CI) methods ignore cross-variable dependencies and model only temporal dynamics, while channel-dependent (CD) methods consider both but typically rely on architectural compromises to mitigate overfitting and computational overhead. We therefore propose Chameleon, a CD state space model (SSM) that enables data-dependent, fine-grained interactions across variables while scaling linearly with the number of variables. By connecting selective SSMs with the Kalman filter, we adapt the latter's measurement-update formulation for cross-variable modeling while preserving the SSM backbone for robust temporal modeling. We further identify favorable inductive biases of GatedDeltaNet for time series, adapt it as our backbone, and improve generalization through additional techniques, including a previously unexplored stochastic perturbation of reversible instance normalization. On strongly dependent ODE and PEMS datasets, Chameleon achieves the best MSE and MAE across all settings, while its CI ablation and prior CD methods incur 61-178% higher MSE on average. Across 28 standard benchmark settings, Chameleon also achieves better MSE and MAE than each baseline in at least 27 and 22 cases, respectively. Training-time, FLOP, and peak-memory analyses on Traffic and ETT further demonstrate competitive efficiency and favorable scalability across different variable counts.

## FlexTSF: A Flexible Forecasting Model for Time Series with Variable Regularities

DevFeed: [FlexTSF: A Flexible Forecasting Model for Time Series with Variable Regularities](<https://devfeed.tech/articles/flextsf-a-flexible-forecasting-model-for-time-series-with-variable-regularities-144849.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2410.23160>)

Author: Jingge Xiao, Yile Chen, Gao Cong, Wolfgang Nejdl, Simon Gottschalk

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

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>)

Tags: [adaptive](<https://devfeed.tech/tags/adaptive.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [attention](<https://devfeed.tech/tags/attention.md>), [continuous](<https://devfeed.tech/tags/continuous.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>), [domains](<https://devfeed.tech/tags/domains.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [ivp](<https://devfeed.tech/tags/ivp.md>), [missing-values](<https://devfeed.tech/tags/missing-values.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [transformer](<https://devfeed.tech/tags/transformer.md>), [zero-shot](<https://devfeed.tech/tags/zero-shot.md>)

### AI overview

FlexTSF targets forecasting across time series with uneven sampling, variable lengths, and missing values. Its continuous-time IVP Patcher handles these irregularities, while a decoder-only Transformer uses normalized timestamps, domain-specific statistical descriptors, and specialized causal self-attention to adapt across domains. The abstract reports experiments on 16 datasets, with FlexTSF outperforming existing models in classic forecasting, zero-shot generalization, and low-resource fine-tuning; it provides no detailed metrics or comparison breakdown.

### Source excerpt

arXiv:2410.23160v3 Announce Type: replace Abstract: Generalizing across time series with irregular temporal structures poses a persistent challenge for universal time series forecasting models. Existing approaches often assume regular sampling or depend heavily on imputation, which limits their applicability in real-world scenarios where irregularities are prevalent due to diverse sensing devices and recording practices. To address this gap, we propose FlexTSF, a flexible and domain-adaptive forecasting model specifically designed for time series with variable temporal regularities. At its foundation lies the IVP Patcher, a continuous-time patching module leveraging Initial Value Problems (IVPs) to inherently handle uneven sampling intervals, variable sequence lengths, and missing values. Built upon a decoder-only Transformer architecture, FlexTSF integrates normalized timestamp inputs and domain-specific statistical descriptors through a specialized causal self-attention mechanism, enabling adaptability across domains. Extensive experiments on 16 datasets demonstrate FlexTSF's effectiveness, significantly outperforming existing models in classic forecasting scenarios, zero-shot generalization, and low-resource fine-tuning conditions.

## The Window Lever: Longer Observation Windows Outweigh the Observed Effects of Data, Capacity, and Architecture in Customer Behavior Forecasting

DevFeed: [The Window Lever: Longer Observation Windows Outweigh the Observed Effects of Data, Capacity, and Architecture in Customer Behavior Forecasting](<https://devfeed.tech/articles/the-window-lever-longer-observation-windows-outweigh-the-observed-effects-of-data-capacity-and-architecture-in-customer-behavior-forecasting-145191.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.26521>)

Author: Tara Sowrirajan, Yanqiao Wang

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

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>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [model architecture](<https://devfeed.tech/topics/model-architecture.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [customer](<https://devfeed.tech/tags/customer.md>), [data](<https://devfeed.tech/tags/data.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [model](<https://devfeed.tech/tags/model.md>), [transformer](<https://devfeed.tech/tags/transformer.md>), [weekly](<https://devfeed.tech/tags/weekly.md>), [window](<https://devfeed.tech/tags/window.md>), [windows](<https://devfeed.tech/tags/windows.md>)

### AI overview

In the reported long-horizon customer-behavior forecasting experiments, extending history from 8 to 16 weeks increased log-space R² by 0.070-0.107 across model families at the 1M tier; matched test cohorts suggest the raw comparison understates the gain. The direction also held on five public datasets. The authors report smaller, unresolved observed effects for data, model capacity, and architecture, and recommend trying longer windows for this class of data. However, longer windows also admit more and colder training customers, and the design does not separate history length from calendar coverage. The authors therefore report a robust positive association, not a measured calendar-independent history effect. The only fully blind registered prediction failed.

### Source excerpt

arXiv:2611.26521v1 Announce Type: new Abstract: In our experiments on long-horizon customer-behavior forecasting, a longer observation window outweighed the observed effects of more data, more capacity and newer architectures. The window is the weeks of history a model sees before the week it forecasts from. Going from an 8- to a 16-week window adds a raw +0.070 to 0.107 log-space $R^2$ in every family at 1M. Matched test cohorts, which score the same customer-weeks at every window, show this raw comparison understates the gain in every family they cover. The direction holds on five public datasets, four spanning over a year (ridge and our transformer on all five). On our internal data, longer windows admit up to 10% more and colder training customers, which matching does not fix. Our design does not separate window length from calendar coverage, because every test anchor (the week a forecast is made from) falls in one eight-week band. On training-period anchors, the weekly ridge probe's gain persists outside the holiday season, keeping about 72% of it. We therefore report a robust positive association with longer windows and do not identify the magnitude of a calendar-independent history effect. Of four internally registered predictions, the only fully blind one failed. An activity-mixture model predicted +0.014 from 24 to 32 weeks; the weekly-aggregate probe returned -0.0004. For the sequence families, the other three levers are about an order of magnitude smaller, bounded and unresolved. Their observed effects are at most 0.006 between the nominal 1M and 10M tiers and 0.0052 among five sequence families at 16 weeks with clocks as built, both single-seed on the full test split, and 0.006 for capacity on three-seed means. These are observed intervals, not equivalence bounds. Our claim is the order-of-magnitude contrast between the window and those bounds. A team building on this class of data should start by lengthening the window.

## Signature-Kernel Based Evaluation Metrics for Robust Probabilistic and Tail-Event Forecasting

DevFeed: [Signature-Kernel Based Evaluation Metrics for Robust Probabilistic and Tail-Event Forecasting](<https://devfeed.tech/articles/signature-kernel-based-evaluation-metrics-for-robust-probabilistic-and-tail-event-forecasting-143234.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2602.10182>)

Author: Benjamin R. Redhead, Thomas L. Lee, Peng Gu, V\\'ictor Elvira, Amos Storkey

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [based](<https://devfeed.tech/tags/based.md>), [batch](<https://devfeed.tech/tags/batch.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [body](<https://devfeed.tech/tags/body.md>), [can](<https://devfeed.tech/tags/can.md>), [capture](<https://devfeed.tech/tags/capture.md>), [code](<https://devfeed.tech/tags/code.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [custom](<https://devfeed.tech/tags/custom.md>), [distribution](<https://devfeed.tech/tags/distribution.md>), [errors](<https://devfeed.tech/tags/errors.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [event](<https://devfeed.tech/tags/event.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [metrics](<https://devfeed.tech/tags/metrics.md>), [probabilistic](<https://devfeed.tech/tags/probabilistic.md>), [robust](<https://devfeed.tech/tags/robust.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>)

### AI overview

The paper proposes censored signature kernel metrics for evaluating probabilistic time-series forecasts in a chosen focus region, such as distribution tails or high-utility events. The method collapses the distribution body to a single pivot, aiming to retain a proper scoring rule while assessing tail fidelity and multivariate correlations that CRPS and Quantile Loss may miss. Benchmarks on time-series foundation models show a ranking for overall distribution capture but no clear winner for tasks such as systemic load prediction, suggesting that distribution fit alone does not establish downstream utility. The authors also introduce the open-source efficient signature kernel (ESK) library, whose custom Triton kernels enable batch computation and reach up to 3.58x speed-up over implementations using SigKernel.

### Source excerpt

arXiv:2602.10182v2 Announce Type: replace Abstract: Standard sample-based metrics like Continuous Ranked Probability Score (CRPS) and Quantile Loss (QL) are frequently used in the evaluation of probabilistic time-series forecasting. However, these metrics fail to capture multivariate correlations or assess the ability to capture distribution tails. These global metrics are insensitive to errors on high-utility events in distributional tails due to over-representation of the distribution's body. To accurately measure distributional fidelity on tail-regions and regions of high utility while still having a proper scoring rule, we introduce a family of censored signature kernel metrics. Our proposed metrics concentrate evaluation of forecasts to a focus region, representing high-utility or distributional tails, by collapsing the body of the distribution to a single pivot. Our benchmarks on time-series foundation models (TSFMs) reveal that while a clear ranking can be formed for distribution capture, there is no clear winner for tasks like systemic load prediction. This indicates that models with a strong ability to capture an overall distribution do not produce forecasts with high downstream utility. To encourage the use of the proposed metrics we open-source the efficient signature kernel (ESK) library. This library facilitates batch computation, with custom triton kernels, achieving a speed-up of up to 3.58$\times$ compared to implementations via the popular SigKernel library. Code for our metrics are available on GitHub.

## CAFE-AQI: Coverage-Aware Day-Ahead AQI Forecasting with Early Warning and Conformal Uncertainty in Bangladesh

DevFeed: [CAFE-AQI: Coverage-Aware Day-Ahead AQI Forecasting with Early Warning and Conformal Uncertainty in Bangladesh](<https://devfeed.tech/articles/cafe-aqi-coverage-aware-day-ahead-aqi-forecasting-with-early-warning-and-conformal-uncertainty-in-bangladesh-143450.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23815>)

Author: Md Rafid Islam, Rafsan Jany, Rafid Bin Zaman, Mohammad Fahim Islam, Hafiz Abdur Rahman

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>)

Tags: [2022](<https://devfeed.tech/tags/2022.md>), [2023](<https://devfeed.tech/tags/2023.md>), [2025](<https://devfeed.tech/tags/2025.md>), [air-quality](<https://devfeed.tech/tags/air-quality.md>), [bangladesh](<https://devfeed.tech/tags/bangladesh.md>), [conditional](<https://devfeed.tech/tags/conditional.md>), [coverage](<https://devfeed.tech/tags/coverage.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [early-warning](<https://devfeed.tech/tags/early-warning.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [spatial](<https://devfeed.tech/tags/spatial.md>), [stat-ap](<https://devfeed.tech/tags/stat-ap.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [testing](<https://devfeed.tech/tags/testing.md>), [training](<https://devfeed.tech/tags/training.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>), [warning](<https://devfeed.tech/tags/warning.md>)

### AI overview

CAFE-AQI forecasts next-day maximum AQI and the probability of unhealthy-or-worse air (AQI >=151), addressing uneven data coverage and duplicate city records. The Bangladesh dataset audit found that 103 city labels represented 61 unique AQI trajectories; identical trajectories were weighted to limit their influence. The framework combines temporal and pollutant features with five-neighbor spatial context, a class-balanced warning head, and risk-conditional conformal calibration. On duplicate-balanced 2025 test data, it achieved MAE 9.76, RMSE 15.04, and R² 0.910, with MAE 3.63 AQI below persistence (95% CI 3.49-3.76) and lower errors across all 61 trajectory groups. Warning performance was F1 0.908 and AUROC 0.982; nominal 90% intervals covered 92.09% of observations. Excluding a spatial block during training yielded MAE 10.66, indicating transfer to unseen locations in this evaluation.

### Source excerpt

arXiv:2611.23815v1 Announce Type: new Abstract: Next-day air quality forecasting often assumes city records are equally complete and independent. However, an audit of a large Bangladesh dataset revealed a sharp coverage expansion after August 2022 and showed that 103 city labels correspond to only 61 unique AQI trajectories. To address these issues, this paper presents CAFE-AQI, a coverage-aware framework for forecasting next-day maximum AQI. It also estimates the probability of Unhealthy-or-worse conditions (AQI >= 151). Data were divided chronologically into training (2022-2023), calibration (2024), and testing (2025). To prevent duplicate records from dominating training and evaluation, identical trajectories were weighted. The framework combines temporal and pollutant features with a five-neighbor spatial context. It also includes a class-balanced warning head and risk-conditional conformal calibration aimed at providing early warnings and prediction intervals. Evaluation on the duplicate-balanced 2025 data gave an MAE of 9.76, an RMSE of 15.04, and an R-squared value of 0.910 for CAFE-AQI. Its MAE was 3.63 AQI lower than persistence (95% CI: 3.49-3.76), with lower errors in each of the 61 trajectory groups. For unhealthy-air warnings, the F1-score was 0.908 and the AUROC was 0.982. The nominal 90% intervals covered 92.09% of observations. When an entire spatial block was excluded from training, the MAE remained 10.66, showing that the lightweight model transferred reasonably well to unseen locations.

## When Do Foundation Models Pay Off? A Break-Even Analysis of Pretrained Time Series Forecasters

DevFeed: [When Do Foundation Models Pay Off? A Break-Even Analysis of Pretrained Time Series Forecasters](<https://devfeed.tech/articles/when-do-foundation-models-pay-off-a-break-even-analysis-of-pretrained-time-series-forecasters-143286.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2607.04919>)

Author: Nicholas Tan Jerome, Frank Simon

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [lora](<https://devfeed.tech/topics/lora.md>), [decision-making](<https://devfeed.tech/topics/decision-making.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>)

Tags: [analysis](<https://devfeed.tech/tags/analysis.md>), [arima](<https://devfeed.tech/tags/arima.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [decision](<https://devfeed.tech/tags/decision.md>), [ets](<https://devfeed.tech/tags/ets.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [fm](<https://devfeed.tech/tags/fm.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [foundation-model](<https://devfeed.tech/tags/foundation-model.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [gpu-infrastructure](<https://devfeed.tech/tags/gpu-infrastructure.md>), [history](<https://devfeed.tech/tags/history.md>), [models](<https://devfeed.tech/tags/models.md>), [series](<https://devfeed.tech/tags/series.md>), [time](<https://devfeed.tech/tags/time.md>), [time-series](<https://devfeed.tech/tags/time-series.md>)

### AI overview

Across 30 datasets, zero-shot and LoRA-tuned time-series foundation models beat classical baselines at every tested training fraction on 16 datasets, while classical methods reached break-even with as little as 21-6,324 samples on six others and 24-22,177 on eight. No classical method dominated. For datasets with fewer than 700 training points and non-negligible seasonality, the study recommends zero-shot forecasting and skipping fine-tuning: at under 20% of history, classical methods lost on four of ten such datasets and could not be fitted on six. LoRA rarely helped and degraded short-series performance by up to 52%. The proposed process checks dataset length and seasonality, then uses a brief 5-10% pilot when needed; at this benchmark scale, simple dataset features did not predict break-even outside the short-seasonal-series rule.

### Source excerpt

arXiv:2607.04919v2 Announce Type: replace Abstract: Deploying a time series foundation model (FM) requires GPU infrastructure and engineering overhead, with no guarantee of improvement over classical forecasting methods. We provide the first systematic break-even analysis of when this investment pays off. Across 30 benchmark datasets, we compare zero-shot and LoRA fine-tuned FMs (Chronos-Bolt, Moirai, Lag-Llama) against classical baselines (Naive, ETS, ARIMA, XGBoost) at six training set sizes from 2% to 100% of available data. FMs outperform classical methods at every evaluated training fraction on 16 of 30 datasets. On 6 datasets, classical methods surpass zero-shot FMs with at most 5% of the training data (21--6,324 samples); on the remaining 8, break-even ranges from 24 to 22,177 samples. No classical method dominates: ARIMA, Naive, ETS and XGBoost are each best on 6--11 datasets. One robust rule requires no model training: if n_train < 700 and seasonality is non-negligible, FM zero-shot is the right low-data choice -- below 20% of the history, classical methods either lose to it (4 of these 10 datasets) or cannot be fitted at all (6) -- and it stays best at full history on 6 of them; fine-tuning should be skipped. Contrary to common practice, LoRA fine-tuning rarely helps and degrades short series by up to 52%. We operationalise these findings as a two-step decision framework -- compute dataset length and seasonality strength, run a brief 5--10% pilot only if needed. Cheap dataset features do not predict the break-even category at this benchmark scale, so outside the rule a pilot remains necessary. Code and benchmark: https://github.com/nicolaisi/fm_breakeven.

## Recurrent Domain Incremental Adaptation of Time-Series Foundation Models

DevFeed: [Recurrent Domain Incremental Adaptation of Time-Series Foundation Models](<https://devfeed.tech/articles/recurrent-domain-incremental-adaptation-of-time-series-foundation-models-142751.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24075>)

Author: Sachith Abeywickrama, Peiliang Gong, Emadeldeen Eldele, Min Wu, Chau Yuen

Published: 2026-11-19T05:00:00Z

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>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>)

Tags: [adapting](<https://devfeed.tech/tags/adapting.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [charts](<https://devfeed.tech/tags/charts.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [eess-sp](<https://devfeed.tech/tags/eess-sp.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [gate](<https://devfeed.tech/tags/gate.md>), [incremental](<https://devfeed.tech/tags/incremental.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [retention](<https://devfeed.tech/tags/retention.md>), [shift](<https://devfeed.tech/tags/shift.md>), [stream](<https://devfeed.tech/tags/stream.md>), [time-series-foundation-models](<https://devfeed.tech/tags/time-series-foundation-models.md>)

### AI overview

PACT addresses incremental adaptation to changing operating conditions by partitioning a frozen time-series foundation model's embedding space into local charts and learning lightweight residual transports only for affected regions. An identity path is retained, and a learned gate selects between it and accumulated transports without requiring condition identity at inference. The encoder, classifier head, and prior transports stay fixed. With 5% labeled calibration data per arriving condition, the authors report average macro-F1 above 0.95 across a four-condition bearing-fault stream; they report similar retention and adaptation on public benchmarks when tasks and label spaces are homogeneous.

### Source excerpt

arXiv:2611.24075v1 Announce Type: new Abstract: Pretrained Time-Series Foundation Models (TSFMs) provide strong representations for downstream classification tasks. Yet adapting a TSFM under changing operating conditions remains challenging. A model must improve on each newly observed condition without sacrificing conditions it has already learned, often from only a small labeled calibration set and without knowing the operating condition at inference time. We find that this difficulty has an important geometric structure. Under an operating-condition shift, the embedding space of a TSFM does not move uniformly: some local regions change substantially, while others remain nearly unchanged. Consequently, a global adaptation can unnecessarily move embeddings that were already correctly classified. To address this, we propose the novel PACT (Piecewise Adaptation via Chart Transport). PACT partitions the frozen source embedding space into local charts and learns lightweight residual transports for the regions affected by each new condition. An explicit identity path is always retained, and a learned gate softly selects among the identity and accumulated transports for each input. The pretrained TSFM encoder, classifier head, and previously learned transports remain fixed. With only 5% labeled calibration data per arriving condition and no condition identity at test time, PACT maintains an average macro-F1 above 0.95 across a four-condition bearing-fault stream. Under homogeneous tasks and label spaces, similar retention-adaptation behavior is observed across public benchmarks. Code is available at https://github.com/Sachithx/PACT.

## Distribution-Free Uncertainty Quantification for Weather Forecasting: Conformal Prediction Under Spatiotemporal Dependence

DevFeed: [Distribution-Free Uncertainty Quantification for Weather Forecasting: Conformal Prediction Under Spatiotemporal Dependence](<https://devfeed.tech/articles/distribution-free-uncertainty-quantification-for-weather-forecasting-conformal-prediction-under-spatiotemporal-dependence-143508.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24329>)

Author: Mariya

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Climate / Weather / Ocean Modeling](<https://devfeed.tech/topics/climate-weather-ocean-modeling.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>)

Tags: [conformal-prediction](<https://devfeed.tech/tags/conformal-prediction.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>), [uncertainty-quantification](<https://devfeed.tech/tags/uncertainty-quantification.md>), [weather](<https://devfeed.tech/tags/weather.md>), [weather-forecasting](<https://devfeed.tech/tags/weather-forecasting.md>)

### AI overview

Per-cell conformal methods achieved near-nominal marginal coverage for gridded 2 m temperature forecasts, but coverage varied sharply across days: the variance in the fraction of covered grid cells was about 3-10 times the independent-cell benchmark, and forecast misses clustered over time. The study compared eight interval procedures on ERA5 at 5.625° resolution using a CNN and ECMWF HRES across eight lead times. An oracle adjustment using each day's mean normalized error removed most excess spatial dispersion, suggesting widespread high-error days drive much of it. A day-scale adjustment based on seasonal terms and verified errors reduced overdispersion without wider intervals, but improved interval score by at most 0.6%; no method performed best on both measures. Results are limited to one variable, a coarse grid, a reanalysis target, and two forecasters.

### Source excerpt

arXiv:2611.24329v1 Announce Type: new Abstract: Conformal prediction gives finite-sample, distribution-free marginal coverage and has begun to be applied to gridded weather forecasts. Marginal coverage concerns the long-run fraction of covered points at a location, not how coverage behaves jointly across space and time on a given day. I benchmark eight interval-construction procedures (Gaussian, pooled and per-cell split conformal, per-season, adaptive conformal inference and rolling windows) for gridded 2 m temperature forecasts at eight lead times from 6 hours to 10 days, using ERA5 at 5.625 degree resolution, a deliberately modest convolutional forecaster and ECMWF's operational HRES, with calibration in 2017-2018 and testing in 2019-2021. Alongside coverage, width and interval score, I report dependence diagnostics: a spatial overdispersion index, quantiles of daily spatial coverage and miss persistence relative to a shuffled-days null. Per-cell methods reach near-nominal marginal coverage, yet the day-to-day variance of the fraction of cells covered is roughly 3 to 10 times what independent cells would give, and misses cluster in time; both effects grow with lead time for the CNN, also appear for HRES, and are shared by all methods. An oracle that divides each day's normalised errors by that day's mean removes 81 to 92 percent of the excess dispersion for the CNN and 93 to 98 percent for HRES, so days of large errors across much of the grid account for most of it. A day-scale predicted from seasonal terms and already-verified errors lowers the overdispersion index without widening intervals but improves interval score by at most 0.6 percent; on HRES from 12 hours to 3 days it has the lowest overdispersion index while ACI or season-conditional calibration has the best interval score, so no method is best on both. The study covers one variable, one coarse grid, a reanalysis target and two forecasters.

## OCTR2A: Real-Time RGB-to-BEV Trajectory Projection and Forecasting at Urban Intersections

DevFeed: [OCTR2A: Real-Time RGB-to-BEV Trajectory Projection and Forecasting at Urban Intersections](<https://devfeed.tech/articles/octr2a-real-time-rgb-to-bev-trajectory-projection-and-forecasting-at-urban-intersections-142782.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24352>)

Author: Emanuele Iacobelli, Simone Teglia, Francesco Pro, Claudia Melis Tonti, Irene Amerini

Published: 2026-11-19T05:00:00Z

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [real-time](<https://devfeed.tech/topics/real-time.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [blender](<https://devfeed.tech/tags/blender.md>), [camera](<https://devfeed.tech/tags/camera.md>), [cameras](<https://devfeed.tech/tags/cameras.md>), [cascade](<https://devfeed.tech/tags/cascade.md>), [class](<https://devfeed.tech/tags/class.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [consumer](<https://devfeed.tech/tags/consumer.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [detection](<https://devfeed.tech/tags/detection.md>), [domains](<https://devfeed.tech/tags/domains.md>), [existing](<https://devfeed.tech/tags/existing.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [projection](<https://devfeed.tech/tags/projection.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [rgb](<https://devfeed.tech/tags/rgb.md>)

### AI overview

OCTR2A reconstructs and forecasts agents' bird's-eye-view trajectories from paired opposite-corner RGB cameras, avoiding reliance at inference time on LiDAR, aerial views, HD maps, or calibrated infrastructure. Its cascade combines per-view detection and class refinement, geometric cross-camera matching, transformer-based BEV projection, and egocentric GRU forecasting. The authors introduce R3ST++, a Blender-based synthetic benchmark built from SIND trajectories, with synchronized views, cross-view correspondences, and BEV ground truth; it also supports site-specific fine-tuning. With 12 observed and 12 predicted frames at 10 Hz, the system reports projection-to-forecasting ADE/FDE of 0.144/0.140 to 0.154/0.241 meters in real time on one consumer GPU. On ground-truth BEV histories, its forecasting module improves ADE/FDE over the strongest compared methods on SIND by 66.7%/41.7%. Transfer experiments indicate R3ST++ pretraining supports adaptation to TUMTraf-V2X roadside domains with different sensing configurations.

### Source excerpt

arXiv:2611.24352v1 Announce Type: cross Abstract: Trajectory forecasting at urban intersections is often studied assuming that reliable bird's-eye-view (BEV) motion histories are available, which in practice requires costly resources such as LiDAR, aerial views, HD maps, or calibrated infrastructure. We address this upstream perception-to-forecasting problem by reconstructing and predicting agents' BEV trajectories only from paired opposite-corner RGB cameras mounted on existing urban infrastructure at variable heights at inference time. We introduce the Opposite-Corner Trajectory Reconstruction and Anticipation Architecture (OCTR2A), a modular pipeline integrating per-view detection with class refinement, geometric cross-camera matching, transformer-based BEV projection, and egocentric GRU forecasting, trained in cascade and mainly supervised with a soft multi-scale spatial-overlap loss. Since no public benchmark provides this sensing configuration, we introduce R3ST++, a Blender-based synthetic benchmark built from SIND trajectories that provides synchronized opposite-corner views, cross-view correspondences, and BEV ground truth, while also serving as a deployment-support simulator for site-specific fine-tuning. Experiments on these datasets show that, with 12 observed and 12 predicted frames at 10\,Hz, OCTR2A achieves projection~-~forecasting ADE/FDE of 0.144/0.140~-~0.154/0.241~meters in real time on a single consumer GPU. Evaluated on ground-truth BEV histories, the forecasting module improves over the strongest state-of-the-art methods on SIND by 66.7\%/41.7\% ADE/FDE. Transfer experiments on TUMTraf-V2X show that R3ST++ pretraining supports adaptation to real-world roadside domains with different sensing configurations.

## Dirichlet-Guided Group Forecasting for Alleviating Over-smoothing in Time Series Forecasting

DevFeed: [Dirichlet-Guided Group Forecasting for Alleviating Over-smoothing in Time Series Forecasting](<https://devfeed.tech/articles/dirichlet-guided-group-forecasting-for-alleviating-over-smoothing-in-time-series-forecasting-143272.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2606.10592>)

Author: Xingyu Zhang, Jingyao Wang, Xin Yu, Zeen Song, Jianqi Zhang, Changwen Zheng, Wenwen Qiang

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>)

Tags: [based](<https://devfeed.tech/tags/based.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [changes](<https://devfeed.tech/tags/changes.md>), [compression](<https://devfeed.tech/tags/compression.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [distributions](<https://devfeed.tech/tags/distributions.md>), [diversity](<https://devfeed.tech/tags/diversity.md>), [evolution](<https://devfeed.tech/tags/evolution.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [framework](<https://devfeed.tech/tags/framework.md>), [future](<https://devfeed.tech/tags/future.md>), [mechanism](<https://devfeed.tech/tags/mechanism.md>), [modal](<https://devfeed.tech/tags/modal.md>), [models](<https://devfeed.tech/tags/models.md>), [multiple](<https://devfeed.tech/tags/multiple.md>), [over](<https://devfeed.tech/tags/over.md>), [series](<https://devfeed.tech/tags/series.md>), [time](<https://devfeed.tech/tags/time.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>)

### AI overview

The abstract attributes forecasting over-smoothing to plausible future dynamics being weakened or averaged under partial observation and single-realization supervision. It proposes Dirichlet-Guided Group Forecasting (DGF), which models multiple mode-conditioned predictive distributions and uncertainty over their selection probabilities. A Dirichlet-guided hierarchical sampling mechanism and reward-based optimization aim to make forecasts accurate, dynamically consistent, and distinct across modes. The abstract reports that experiments on real-world benchmarks reduced over-smoothing while improving accuracy, diversity, and dynamical consistency; it provides no benchmark-specific results in this excerpt.

### Source excerpt

arXiv:2606.10592v2 Announce Type: replace Abstract: Time series forecasting often suffers from over-smoothing, especially when future dynamics are multi-modal. Forecasts may follow the coarse trend of the observed future, but fail to preserve sharp changes, oscillations, turning points, and regime transitions that define plausible dynamic evolution. In this work, we revisit over-smoothing from the perspective of latent dynamical mode compression: under partial observation and single-realization supervision, multiple plausible future modes can be weakened, merged, or averaged during forecasting. Based on this view, we propose Dirichlet-Guided Group Forecasting (DGF), a mode-preserving forecasting framework that explicitly models multiple mode-conditioned predictive distributions and uncertainty over their selection probabilities. DGF uses a Dirichlet-guided hierarchical sampling mechanism and reward-based optimization to encourage forecasts that are accurate, dynamically consistent, and mode-distinct. Extensive experiments on real-world forecasting benchmarks show that DGF reduces over-smoothing while improving forecasting accuracy, diversity, and dynamical consistency. Our code is publicly available at https://github.com/Zh-XY22/DGF

## From Arraignment to Resolution: Transformer Models of Court Cases

DevFeed: [From Arraignment to Resolution: Transformer Models of Court Cases](<https://devfeed.tech/articles/from-arraignment-to-resolution-transformer-models-of-court-cases-143402.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23518>)

Author: Kory Rosen, Aron Culotta

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [Data analysis](<https://devfeed.tech/topics/data-analysis.md>)

Tags: [cases](<https://devfeed.tech/tags/cases.md>), [court](<https://devfeed.tech/tags/court.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-cy](<https://devfeed.tech/tags/cs-cy.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [transformer](<https://devfeed.tech/tags/transformer.md>), [transformer-models](<https://devfeed.tech/tags/transformer-models.md>)

### AI overview

The paper models court cases as event sequences to forecast the next proceeding, such as charges accepted, bond set, or a plea deal. Its transformer combines clerk-note text with structured information about defendants, prior charges, and courtroom. On thousands of cases from publicly available records in one mid-sized U.S. city, it reached 0.57 micro-F1 on a 20-class multilabel task, outperforming several baselines and performing comparably to a GRU with the same inputs. The authors present the model as a way to support oversight and help observers identify emerging issues in that jurisdiction; the excerpt does not establish performance or applicability elsewhere.

### Source excerpt

arXiv:2611.23518v1 Announce Type: cross Abstract: The criminal legal system is composed of many complex decision points, such as which charges are pursued, how bond is determined, plea deal negotiations, and sentencing. Given the profound impact of such decisions, there is a growing need to increase the transparency within this system. While prior machine learning approaches have focused on modeling single decisions (e.g., recidivism), this paper proposes a transformer-based time series model to forecast the outcome of the next proceeding in a case (e.g., charges accepted, bond set, plea deal made, etc.). By modeling a case as an event sequence, we enable a more nuanced analysis to support better oversight and understanding of decision making. We conduct an empirical analysis of thousands of cases using a newly constructed dataset derived from publicly available court records of a mid-sized U.S. city. We fit models that combine textual data from clerk notes with structured information about the defendant, prior charges, and courtroom. Our transformer model achieves $.57$ micro-F1 on a 20-class multi-label task, a significant improvement over several baselines and comparable to a GRU trained on the same input. We demonstrate how the model can be used to provide insights into decision-making and help court observers identify emerging issues within the jurisdiction studied.

## TRESS: A Fixed-Size History Representation for Irregular Time Series Forecasting

DevFeed: [TRESS: A Fixed-Size History Representation for Irregular Time Series Forecasting](<https://devfeed.tech/articles/tress-a-fixed-size-history-representation-for-irregular-time-series-forecasting-143500.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24257>)

Author: Shusen Ma, Yangchun Xie, Xudong Zuo, Haizhou Liang, Yun-Bo Zhao

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

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

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Time-sequences](<https://devfeed.tech/topics/time-sequences.md>), [Temporal data](<https://devfeed.tech/topics/temporal-data.md>)

Tags: [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [cpu-inference](<https://devfeed.tech/tags/cpu-inference.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [history](<https://devfeed.tech/tags/history.md>), [size](<https://devfeed.tech/tags/size.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>)

### AI overview

TRESS represents irregular time series with a fixed-size history: it keeps each variable's latest observations at their original timestamps and summarizes the full history, so representation size stays constant as records grow. Its query-conditioned event lifting weights deviations from the latest value by observation age and forecast horizon; a channel-specific decoder combines these features with historical statistics and a cross-variable summary to predict a correction. With 8-16 retained observations per variable, TRESS achieved the lowest mean test MSE and MAE among methods compared on all four benchmarks. Reusing one encoded history for 32 successive query times reduced CPU inference time by 2.12-2.60x across those datasets.

### Source excerpt

arXiv:2611.24257v1 Announce Type: new Abstract: A fixed number of recent observations may span minutes for one variable and days for another. Earlier measurements can still reveal whether the latest value is unusual for that variable. We propose TRESS (Timestamped Recent Events and Statistical Summaries), a forecasting model that retains the latest observations at their original times and summarises the full history of each variable. For a given set of variables and event budget, this representation stays the same size as the record grows. Its query-conditioned event lifting module weights deviations from the latest value according to the age of each observation and the forecast horizon. A channel-specific decoder combines the resulting features with historical statistics and a summary shared across variables to predict a correction to the latest value. Retaining 8 to 16 observations per variable, TRESS records the lowest mean test MSE and MAE among the methods compared on all four benchmarks. Serving 32 successive query times from one encoded history reduces CPU inference time by a factor of 2.12--2.60 across the four datasets. The code is at https://github.com/SSMa913/NICLab-TRESS.

## Time-Series Models in the Streamhouse: A New Frontier and Use Cases

DevFeed: [Time-Series Models in the Streamhouse: A New Frontier and Use Cases](<https://devfeed.tech/articles/time-series-models-in-the-streamhouse-a-new-frontier-and-use-cases-143009.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/ibm-research/time-series-powered-streamhouse>)

Author: Nicholas Fuller; Sean Falconer; Ayhan Sebin; Pankaj Dayama; Roman Vaculin; Bilge Zeren Aksu; Julian Payne; Mountu Jinwala; Wesley M Gifford; Chandra Reddy; Vijay Ekambaram; Arindam Jati; Robert Metzge

Published: 2026-11-18T15:00:23Z

Content type: release

Language: en

Sources: [Hugging Face](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [streaming data analytics](<https://devfeed.tech/topics/streaming-data-analytics.md>), [Event-Streaming](<https://devfeed.tech/topics/event-streaming.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Confluent Platform](<https://devfeed.tech/topics/confluent-platform.md>)

Tags: [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.md>), [business](<https://devfeed.tech/tags/business.md>), [can](<https://devfeed.tech/tags/can.md>), [cases](<https://devfeed.tech/tags/cases.md>), [change](<https://devfeed.tech/tags/change.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [confluent](<https://devfeed.tech/tags/confluent.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cpus](<https://devfeed.tech/tags/cpus.md>), [detection](<https://devfeed.tech/tags/detection.md>), [downloads](<https://devfeed.tech/tags/downloads.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [form](<https://devfeed.tech/tags/form.md>), [frontier](<https://devfeed.tech/tags/frontier.md>), [help](<https://devfeed.tech/tags/help.md>), [models](<https://devfeed.tech/tags/models.md>), [series](<https://devfeed.tech/tags/series.md>), [time](<https://devfeed.tech/tags/time.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [use-cases](<https://devfeed.tech/tags/use-cases.md>)

### AI overview

IBM time-series foundation models are available as native Confluent functions in managed Confluent Cloud and on-premises Confluent Platform for Apache Flink, running on CPUs. Flink can apply forecasting, anomaly detection, and time-series representations to streaming data with SQL, putting results alongside current events. Pretrained models offer a starting point across many series; workflows can incorporate recent observations, business rules, and human review. Examples include forecasting production quality to evaluate setpoints, flagging unusual store sales against a chosen confidence level, refreshing store-level demand forecasts to inform reorder points, adding individual spending patterns to existing payment checks, and pacing trades from short-term volume forecasts. Similarity search with TSPulse requires a separately built embedding pipeline and index, and can help investigate patterns but cannot establish root cause. These scenarios are illustrative; outcomes depend on model requirements, pipeline configuration, and application logic.

### Source excerpt

**Real-time forecasting, anomaly detection, and optimization, powered by IBM time-series models with more than 50 million downloads. Now available natively on Confluent, in the cloud and on premises, running on CPUs.** Every business generates a continuous record of change: a machine reading, a payment, a sale, an order, or a change in inventory. These observations form time series, and they carry signals that can help teams anticipate what happens next.

## Pseudo-Label-Triggered Retraining from Forecast Errors for Online Time Series Forecasting

DevFeed: [Pseudo-Label-Triggered Retraining from Forecast Errors for Online Time Series Forecasting](<https://devfeed.tech/articles/pseudo-label-triggered-retraining-from-forecast-errors-for-online-time-series-forecasting-139826.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.39789>)

Author: Yeryeong Kwak, Yoo-Min Jung, Jonghun Park

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

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>)

Tags: [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>), [errors](<https://devfeed.tech/tags/errors.md>), [forecast](<https://devfeed.tech/tags/forecast.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [online](<https://devfeed.tech/tags/online.md>)

### AI overview

PILOT learns when to retrain an online time-series forecasting model from realized forecast-error patterns. Because ground-truth retraining labels are unavailable, it creates pseudo-labels from future error increases and trains a lightweight scorer using observed error states. At deployment, the trigger relies only on completed forecast errors and can plug into forecasting backbones without changing their architecture. The abstract reports evaluation on eight multivariate forecasting benchmarks with DLinear, iTransformer, and TimesNet; across all three, PILOT achieved the best average rank among the evaluated retraining policies while maintaining a favorable performance-efficiency trade-off. The excerpt does not provide numerical results or details of the compared policies.

### Source excerpt

arXiv:2609.39789v2 Announce Type: replace-cross Abstract: Real-world time series forecasting systems operate under non-stationary data streams, where forecasting performance may degrade over time. Although retraining can recover the performance, it incurs non-trivial computational and operational costs. Under limited deployment resources, the key challenge is therefore not only how to retrain but also when to retrain. While existing retraining policies often rely on indirect indicators such as drift alarms or model staleness, we instead use realized forecast errors as direct deployment feedback. In this paper, we propose PILOT (Pseudo-label-Informed Learned Online Trigger), an online retraining framework that learns when to retrain from forecast-error dynamics. Since ground-truth retraining labels are unavailable, PILOT constructs a pseudo-label from future increases in forecast error and trains a lightweight scorer to predict it from observed error states. At deployment, PILOT uses only completed forecast errors and serves as a plug-in module for arbitrary forecasting backbones without architectural modification. We evaluate PILOT under standard multivariate forecasting settings across eight benchmarks with three representative backbones---DLinear, iTransformer, and TimesNet. Across all three backbones, PILOT achieves state-of-the-art average-rank performance among retraining policies while maintaining a favorable performance--efficiency trade-off.

## Reason over time-series data with OpenTSLM on Amazon SageMaker JumpStart

DevFeed: [Reason over time-series data with OpenTSLM on Amazon SageMaker JumpStart](<https://devfeed.tech/articles/reason-over-time-series-data-with-opentslm-on-amazon-sagemaker-jumpstart-138239.md>)

Original publisher: [Read original article](<https://aws.amazon.com/blogs/machine-learning/reason-over-time-series-data-with-opentslm-on-amazon-sagemaker-jumpstart/>)

Author: Martin Maritsch

Published: 2026-11-17T18:06:44Z

Content type: article

Language: en

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

Topics: [Amazon SageMaker JumpStart](<https://devfeed.tech/topics/amazon-sagemaker-jumpstart.md>), [Temporal data](<https://devfeed.tech/topics/temporal-data.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [Amazon SageMaker](<https://devfeed.tech/topics/amazon-sagemaker.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [Data pipelines](<https://devfeed.tech/topics/data-pipelines.md>)

Tags: [advanced-300](<https://devfeed.tech/tags/advanced-300.md>), [amazon-sagemaker](<https://devfeed.tech/tags/amazon-sagemaker.md>), [amazon-sagemaker-jumpstart](<https://devfeed.tech/tags/amazon-sagemaker-jumpstart.md>), [deploy](<https://devfeed.tech/tags/deploy.md>), [it-support](<https://devfeed.tech/tags/it-support.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [proactive](<https://devfeed.tech/tags/proactive.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [technical-how-to](<https://devfeed.tech/tags/technical-how-to.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [time-series-data](<https://devfeed.tech/tags/time-series-data.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

OpenTSLM-Flamingo, a 5.24-billion-parameter model combining Gemma 3 4B with a time-series encoder and cross-attention, is available through SageMaker JumpStart for deployment to a real-time endpoint in an AWS environment. Requests pair numerical series with questions and optional context; each series can contain up to 32,768 values. A request supports up to four instances and 32 series per instance, but a 60-second invocation limit and sequential processing constrain long answers or batches. The text also recommends testing input windows because evaluated time-series models can lose event-retrieval accuracy on longer histories. For adaptation, a SageMaker training job loads ARFBench through TimeNet, fine-tunes selected model components while freezing the rest of Gemma 3, and stores the resulting artifact in S3. In the reported run, held-out ARFBench accuracy increased from 5.0% to 45.9%; the article says this is an example result, not a model comparison or official leaderboard score. It recommends evaluating representative cases before deciding whether to improve context or fine-tune.

### Source excerpt

OpenTSLM, an open source family of time-series language models, is now available in Amazon SageMaker JumpStart. See how Nexthink and Aionic Labs apply it to proactive IT support, then deploy OpenTSLM to a real-time endpoint and fine-tune it on your own data with TimeNet and a SageMaker training job.

## SLICES: A Benchmark and Evaluation Framework for Self-Supervised Learning on Sparse ICU Time Series

DevFeed: [SLICES: A Benchmark and Evaluation Framework for Self-Supervised Learning on Sparse ICU Time Series](<https://devfeed.tech/articles/slices-a-benchmark-and-evaluation-framework-for-self-supervised-learning-on-sparse-icu-time-series-138799.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.18868>)

Author: Hannes Ill, Dimitrios Proios, Leo Anthony Celi, Peter J. Sch\\"uffler

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

Content type: article

Language: en

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

Topics: [Self-supervised learning](<https://devfeed.tech/topics/self-supervised-learning.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [reproducibility](<https://devfeed.tech/topics/reproducibility.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [classification](<https://devfeed.tech/tags/classification.md>), [code](<https://devfeed.tech/tags/code.md>), [comparisons](<https://devfeed.tech/tags/comparisons.md>), [contrastive-learning](<https://devfeed.tech/tags/contrastive-learning.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [data-processing](<https://devfeed.tech/tags/data-processing.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [error](<https://devfeed.tech/tags/error.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [framework](<https://devfeed.tech/tags/framework.md>), [hospital](<https://devfeed.tech/tags/hospital.md>), [metric](<https://devfeed.tech/tags/metric.md>), [objectives](<https://devfeed.tech/tags/objectives.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [self-supervised-learning](<https://devfeed.tech/tags/self-supervised-learning.md>), [slices](<https://devfeed.tech/tags/slices.md>), [time-series](<https://devfeed.tech/tags/time-series.md>)

### AI overview

SLICES standardizes comparisons of self-supervised learning for sparse ICU time series by fixing 24-hour value and observation-mask inputs, a Transformer encoder, patient-level splits, four endpoints, two adaptation protocols, and five matched seeds across MIMIC-IV, eICU-CRD, and their pooled union. Of the tested SSL objectives, masked autoencoding (MAE) leads mean AUPRC in all nine frozen classification comparisons and seven of nine fine-tuning comparisons. Fine-tuned MAE leads hospital mortality at full labels and beats a same-architecture supervised Transformer at every label fraction; SSL fine-tuning also has the lowest length-of-stay error. Supervised baselines remain strongest for 24-hour mortality (GRU-D) and acute kidney injury (XGBoost). Rankings vary with adaptation protocol, comparator set, and probe-loss alignment. A two-seed MIMIC-IV rerun stays within two across-seed standard deviations in all 48 downstream runs. Generalization to unseen hospitals was not tested.

### Source excerpt

arXiv:2611.18868v1 Announce Type: new Abstract: Self-supervised learning (SSL) results for intensive care unit (ICU) time series are hard to compare when data processing, encoders, and adaptation vary together. We present SLICES, a benchmark and evaluation framework for comparing and extending SSL methods under a shared clinical setup. It fixes 24-hour value and observation-mask inputs, a Transformer encoder, patient-level splits, four endpoints, two adaptation protocols, and five matched seeds across MIMIC-IV, eICU-CRD, and their pooled union. Masked autoencoding (MAE), joint-embedding prediction (JEPA), and contrastive learning provide reference implementations. Among these objectives, MAE leads mean AUPRC in all nine frozen classification comparisons, also under converged probes, and in seven of nine fine-tuning comparisons. Fine-tuned MAE leads hospital mortality at full labels and scores higher than a supervised Transformer of the same architecture at every label fraction, and SSL fine-tuning gives the lowest length-of-stay error. Supervised baselines remain strong elsewhere: GRU-D leads 24-hour mortality and XGBoost leads acute kidney injury in all three regimes. Rankings depend on the adaptation protocol, the comparator set, and whether the probe loss matches the metric. A two-seed MIMIC-IV rerun with the maintained code stays within two across-seed SDs of the reference in all 48 downstream runs. The code repository supports new objectives, shared downstream evaluation, and regeneration of the reference results, so new SSL methods can be compared with these references under the same conditions. Generalization to unseen hospitals is not tested here and is the natural next extension.

## Severe but Harmless: Stress-Testing Shift Severity for Frozen Time-Series Forecasters

DevFeed: [Severe but Harmless: Stress-Testing Shift Severity for Frozen Time-Series Forecasters](<https://devfeed.tech/articles/severe-but-harmless-stress-testing-shift-severity-for-frozen-time-series-forecasters-137717.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21562>)

Author: Korede R. Bishi, Dominic Okonkwo, Adetayo Okunoye

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>)

Tags: [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.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>), [distribution](<https://devfeed.tech/tags/distribution.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [model-degradation](<https://devfeed.tech/tags/model-degradation.md>), [shift](<https://devfeed.tech/tags/shift.md>), [stress](<https://devfeed.tech/tags/stress.md>), [testing](<https://devfeed.tech/tags/testing.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [time-series-foundation-models](<https://devfeed.tech/tags/time-series-foundation-models.md>), [traffic](<https://devfeed.tech/tags/traffic.md>), [zero-shot](<https://devfeed.tech/tags/zero-shot.md>)

### AI overview

In tests of instance-normalized, frozen time-series foundation models, three established statistical shift-severity measures did not reliably predict model degradation. Across nine regime shifts from three standard traffic datasets, the regimes scored as most severe were not those causing the greatest degradation; degradation reached 29% in the lowest-scoring regime, and most cases showed no positive relationship between severity scores and degradation. The authors use equivariance-based stress testing to explain how measured shift severity can become decoupled from model performance. The supplied text is an abstract and does not provide further methodological detail.

### Source excerpt

arXiv:2611.21562v1 Announce Type: cross Abstract: Zero-shot time-series foundation models (TSFMs) have become attractive alternatives to task-specific models because they handle diverse downstream tasks, including anomaly detection and forecasting, without per-series training. However, time-series data undergo distribution shifts, and the dissimilarity between two regimes is commonly scored with statistical measures such as Wasserstein-1, where a higher score is read as a more severe shift. In systematic experiments using three established severity measures on nine regime shifts from three standard traffic datasets, we show that these measures do not reliably predict degradation for instance-normalized TSFMs. The regimes rated most severe are not the regimes that degrade the models most, and degradation reaches 29% in the lowest-scoring regime. In most cases, the audit shows no positive relationship between measured severity and model degradation. We provide an equivariance-based stress-testing analysis that explains why severity readings can become decoupled from model degradation.

## Conformal Expert Pools: Training-Free Probabilistic Time-Series Forecasting Beyond Seasonality

DevFeed: [Conformal Expert Pools: Training-Free Probabilistic Time-Series Forecasting Beyond Seasonality](<https://devfeed.tech/articles/conformal-expert-pools-training-free-probabilistic-time-series-forecasting-beyond-seasonality-137672.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21267>)

Author: Valery Manokhin

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: [Time Series](<https://devfeed.tech/topics/time-series.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [amazon](<https://devfeed.tech/tags/amazon.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [beats](<https://devfeed.tech/tags/beats.md>), [bench](<https://devfeed.tech/tags/bench.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [better](<https://devfeed.tech/tags/better.md>), [builds](<https://devfeed.tech/tags/builds.md>), [cold-start](<https://devfeed.tech/tags/cold-start.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [free](<https://devfeed.tech/tags/free.md>), [probabilistic](<https://devfeed.tech/tags/probabilistic.md>), [series](<https://devfeed.tech/tags/series.md>), [time](<https://devfeed.tech/tags/time.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>), [training](<https://devfeed.tech/tags/training.md>)

### AI overview

CEP creates probabilistic time-series forecasts without training across series: it combines a pool of simple forecasters, uses their past errors to form predictive distributions, and adapts the pool and errors to each series. On 232 datasets, it beat CSP on most datasets across CRPS, MSIS, and Winkler score; typical CRPS and MSIS were 14% and 23% lower, and 95% interval coverage rose from 90.1% to 92.2%. CEP also had lower CRPS than DeepNPTS, NPTS, and SeasonalNPTS on most datasets, with gains reported on additional dataset families and files. In a separate 228-dataset, multi-step comparison, CSP beat Laplace on most datasets by CRPS and interval score; CEP beat Laplace on most by CRPS and MSIS, with intervals about 40% narrower and runtime 271 times faster. These are reported benchmark results, not evidence of performance on every series.

### Source excerpt

arXiv:2611.21267v1 Announce Type: cross Abstract: Conformal Seasonal Pools (CSP) showed that a forecaster with no training at all, built only from past values at the same point in the season and from past forecast errors, significantly outperformed DeepNPTS, Amazon's learned non-parametric forecaster. We propose Conformal Expert Pools (CEP), a training-free successor. CEP replaces the single seasonal-naive forecast by a pool of simple forecasters, builds its predictive distribution from their past forecast errors, and adapts both the pool and the errors to the series being forecast. No model is trained and no parameter is learned across series. On the CSP benchmark of 232 datasets and 24,502 forecast windows, CEP is better than CSP on 78% of datasets by CRPS, on 84% by MSIS and on 83% by Winkler score; on a typical dataset its CRPS is 14% lower and its MSIS 23% lower (both p < 1e-22, paired Wilcoxon test), and coverage of 95% intervals rises from 90.1% to 92.2%. CEP has a lower CRPS than DeepNPTS on 88% of datasets, than NPTS on 89% and than SeasonalNPTS on 83%. The improvement persists on dataset families outside the development set and on 36 additional fev-bench files. We also test the published claims that Laplace, an online ensemble forecaster, "beats (almost) everything" and "beats" CSP. On 228 datasets at multi-step horizons the claims do not hold: CSP outperforms Laplace on 60% of datasets by CRPS and on 64% by interval score, and Laplace ranks sixth of seven methods, with the widest intervals of all (69% wider than CEP's). Its point forecast is biased low and on 54% of datasets is less accurate than seasonal naive, and from a cold start it takes about 7,700 times longer than CSP. CEP is better than Laplace on 82% of datasets by CRPS and on 80% by MSIS, with intervals about 40% narrower, and runs 271 times faster.

## Look Back, Correct Forward: Error Feedback Calibration for Time Series Forecasting

DevFeed: [Look Back, Correct Forward: Error Feedback Calibration for Time Series Forecasting](<https://devfeed.tech/articles/look-back-correct-forward-error-feedback-calibration-for-time-series-forecasting-137295.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.22872>)

Author: Haonan Yang, Yulin He, Zhuo Li, Jing Zhao, Huibin Tan

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

Content type: article

Language: en

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

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [Temporal data](<https://devfeed.tech/topics/temporal-data.md>)

Tags: [controls](<https://devfeed.tech/tags/controls.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [error](<https://devfeed.tech/tags/error.md>), [errors](<https://devfeed.tech/tags/errors.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [feedback](<https://devfeed.tech/tags/feedback.md>), [forecast](<https://devfeed.tech/tags/forecast.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [framework](<https://devfeed.tech/tags/framework.md>), [future](<https://devfeed.tech/tags/future.md>), [learning](<https://devfeed.tech/tags/learning.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [out-of-sample](<https://devfeed.tech/tags/out-of-sample.md>), [performance](<https://devfeed.tech/tags/performance.md>), [post](<https://devfeed.tech/tags/post.md>), [predictions](<https://devfeed.tech/tags/predictions.md>), [series](<https://devfeed.tech/tags/series.md>), [time](<https://devfeed.tech/tags/time.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [time-series-forecasting](<https://devfeed.tech/tags/time-series-forecasting.md>)

### AI overview

Error Feedback Calibration (EFC) uses revealed errors from overlapping past forecasts to correct future predictions while leaving the forecasting model frozen. It builds a Causal Error Diagonal (CED) to represent recent failure behavior, then uses Cumulative Error Transport (CET) to estimate future residuals; temporal out-of-sample validation controls the correction strength. Temporal Residual Learning (TRL) predicts only the residual structure left after CET, using recent CED trajectories. The abstract reports improvements across forecasting backbones and datasets while substantially controlling harmful overcorrection; it does not provide specific results or implementation details.

### Source excerpt

arXiv:2610.22872v2 Announce Type: replace Abstract: Time series forecasting models are increasingly deployed in continuously evolving environments. After a forecast is issued, future observations gradually arrive and reveal how the model has failed. Existing methods mainly treat such observations as supervision for parameter updates. However, it remains unclear whether verified forecasting errors can themselves provide causal feedback for correcting future predictions. To address this, we introduce Error Feedback Calibration (EFC), a fully test-time framework for deterministic post-hoc correction of time series forecasting. EFC constructs a Causal Error Diagonal (CED) from revealed errors of overlapping historical forecasts, providing a fresh view of the frozen forecaster's recent failure behavior. Then Cumulative Error Transport (CET) maps this error state to future residuals using cumulative estimation, while temporal out-of-sample validation controls the correction strength. Temporal Residual Learning (TRL) further predicts only the residual structure left after CET from recent CED trajectories, rather than relearning the entire forecasting error. Extensive experiments across forecasting backbones and datasets show that EFC broadly improves forecasting performance while substantially controlling harmful over correction.

## Residual-Tail-Informed Flow Matching for Time Series Anomaly Detection

DevFeed: [Residual-Tail-Informed Flow Matching for Time Series Anomaly Detection](<https://devfeed.tech/articles/residual-tail-informed-flow-matching-for-time-series-anomaly-detection-137563.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.20334>)

Author: Shyam Sundar Saravanan, Tie Luo, Sanjay Madria

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [AI Models](<https://devfeed.tech/topics/ai-models.md>), [VAE](<https://devfeed.tech/topics/vae.md>), [Temporal data](<https://devfeed.tech/topics/temporal-data.md>)

Tags: [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.md>), [autoencoder](<https://devfeed.tech/tags/autoencoder.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [flow-matching](<https://devfeed.tech/tags/flow-matching.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [time-series](<https://devfeed.tech/tags/time-series.md>)

### AI overview

RT-FM targets false alarms on rare but valid time-series regimes that data-space flow matching may score like true anomalies. It uses a temporal variational autoencoder to estimate residual latent tailness, then turns that estimate into bounded, warmup-annealed weights for training an unsupervised data-space flow-matching model. Across seven TSAD benchmarks, the authors report better rare-normal coverage, fewer false alarms on those windows, and improved aggregate detection while preserving strong anomaly discrimination.

### Source excerpt

arXiv:2611.20334v1 Announce Type: cross Abstract: Time series anomaly detection (TSAD) models are often trained on nominal data using average-risk objectives, which emphasize dominant normal patterns. However, real time series can contain rare but valid nominal regimes that are underrepresented in the training distribution. We show that data-space flow matching can over-penalize such rare-normal windows and assign them anomaly scores comparable to those of true anomalies. To address this, we propose Residual-Tail-Informed Flow Matching (RT-FM), an unsupervised framework that uses a temporal variational autoencoder to estimate residual latent tailness and converts it into bounded, warmup-annealed weights for training a data-space flow matching model. Experiments on seven TSAD benchmarks show that RT-FM improves rare-normal coverage while preserving strong anomaly discrimination, substantially reducing false alarms on such windows and resulting in improved aggregate detection performance. The code is available at: https://github.com/trustAIoT/RT-FM_time-series-anomaly

## MSAR: Next-Scale Autoregressive Forecasting for Time Series via Modular Multi-Scale Decoupling

DevFeed: [MSAR: Next-Scale Autoregressive Forecasting for Time Series via Modular Multi-Scale Decoupling](<https://devfeed.tech/articles/msar-next-scale-autoregressive-forecasting-for-time-series-via-modular-multi-scale-decoupling-139089.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.20714>)

Author: Fanda Fan, Kuiye Ding, Liming Mao, Bingsong Wang, Yao Wang, Xiaorui Wang, Ruijie Jian, Zhipeng Liu, Luqi Gong, Zhenghua Lu, Chunjie Luo, Jianfeng Zhan

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

Content type: article

Language: en

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

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [autoregressive](<https://devfeed.tech/tags/autoregressive.md>), [cnns](<https://devfeed.tech/tags/cnns.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [decoupling](<https://devfeed.tech/tags/decoupling.md>), [deep](<https://devfeed.tech/tags/deep.md>), [design](<https://devfeed.tech/tags/design.md>), [effectiveness](<https://devfeed.tech/tags/effectiveness.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [energy](<https://devfeed.tech/tags/energy.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [finance](<https://devfeed.tech/tags/finance.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [modular](<https://devfeed.tech/tags/modular.md>), [next](<https://devfeed.tech/tags/next.md>), [scale](<https://devfeed.tech/tags/scale.md>), [series](<https://devfeed.tech/tags/series.md>), [time](<https://devfeed.tech/tags/time.md>), [time-series](<https://devfeed.tech/tags/time-series.md>)

### AI overview

MSAR forecasts time series progressively across multiple temporal resolutions, aligning inputs and outputs at each scale. Its scale-wise autoregression lets coarse-scale predictions guide finer-scale forecasts, while its modular design can integrate with CNN, MLP, and Transformer backbones. The abstract reports consistent gains in accuracy and inference efficiency across datasets and forecasting models; it provides no quantitative results or details of the experiments.

### Source excerpt

arXiv:2611.20714v1 Announce Type: new Abstract: Time series forecasting underpins critical applications in finance, energy, healthcare, and transportation. Although deep models have achieved strong results, most adopt single-scale modeling or restrict multiscale processing to the input side, causing a misalignment between multiscale inputs and single-scale outputs and limiting predictive power. We introduce the Modular Scale-wise Autoregressive Framework (MSAR), a model-agnostic design that forecasts progressively across multiple temporal resolutions. MSAR offers three advantages: (1) scale-wise aligned modeling, which disentangles heterogeneous temporal patterns by aligning inputs and outputs at each scale; (2) scale-wise autoregression, where coarse-scale predictions guide finer-scale forecasting through hierarchical information flow; and (3) a modular architecture, enabling seamless integration with diverse backbones such as CNNs, MLPs, and Transformers. Extensive experiments across a broad set of datasets and forecasting models demonstrate that MSAR achieves consistent improvements in both accuracy and inference efficiency, validating the effectiveness of scale-aligned autoregression for multiscale time series forecasting. The code is available at: https://github.com/Oryxo/MSAR.

## Breaking feature collapse in self-supervised time series encoders

DevFeed: [Breaking feature collapse in self-supervised time series encoders](<https://devfeed.tech/articles/breaking-feature-collapse-in-self-supervised-time-series-encoders-137639.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.20981>)

Author: \\"Ozg\\"un Turgut, Daniel Rueckert

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: [Time Series](<https://devfeed.tech/topics/time-series.md>), [Temporal data](<https://devfeed.tech/topics/temporal-data.md>)

Tags: [code](<https://devfeed.tech/tags/code.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data-modelling](<https://devfeed.tech/tags/data-modelling.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [distillation](<https://devfeed.tech/tags/distillation.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [rank](<https://devfeed.tech/tags/rank.md>), [scale](<https://devfeed.tech/tags/scale.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [time-series](<https://devfeed.tech/tags/time-series.md>)

### AI overview

The authors attribute poor scaling in general-purpose time-series encoders to partial collapse of output tokens under masked data modeling: input-space reconstruction alone does not constrain redundant feature dimensions. OTISv2 adds feature-space self-distillation against an exponential-moving-average teacher and a KoLeo regularizer alongside reconstruction. In matched training, it reportedly lowers token cosine similarity, raises effective rank, and scales monotonically from 7.1M to 116M parameters, while its predecessor declines. Across 143 uni- and multivariate datasets, it matches or surpasses existing general-purpose encoders. The excerpt says code and pretrained weights are released.

### Source excerpt

arXiv:2611.20981v1 Announce Type: cross Abstract: General-purpose time series encoders promise a single backbone that generalises across tasks and domains; however, they currently do not scale. We trace this down to a critical point of failure: under the predominant masked data modelling supervision, the encoder's output tokens partially collapse. This is the natural consequence of supervising encoders only in the input space, which leaves no constraint against redundant token dimensions in the feature space. We introduce OTISv2, which suppresses the collapse by complementing the reconstruction objective in input space with two objectives in the feature space: a self-distillation loss against an exponential-moving-average teacher, and a KoLeo regulariser. Under matched training, the resulting encoder reduces the cosine similarity and lifts the effective rank of its output tokens relative to its predecessor, and it scales monotonically from 7.1 M to 116 M parameters where its predecessor declines. It matches or surpasses existing general-purpose time series encoders across 143 uni- and multi-variate datasets. To foster further research, we release our code and pre-trained weights.

## Auditing World-Model Explanations in Retail Forecasting: Faithfulness, Accuracy and Price Response

DevFeed: [Auditing World-Model Explanations in Retail Forecasting: Faithfulness, Accuracy and Price Response](<https://devfeed.tech/articles/auditing-world-model-explanations-in-retail-forecasting-faithfulness-accuracy-and-price-response-137614.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.20811>)

Author: Kathik Iyer

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: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [enhancing AI model accuracy](<https://devfeed.tech/topics/enhancing-ai-model-accuracy.md>)

Tags: [auditing](<https://devfeed.tech/tags/auditing.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [retail](<https://devfeed.tech/tags/retail.md>), [world-model](<https://devfeed.tech/tags/world-model.md>)

### AI overview

Passing explanation checks did not ensure accurate retail forecasts or price-response claims. In generated store data, a world model passed checks for construction, stability, price direction and agreement with sales and LightGBM, yet overstated promotion lift by 1.33-1.40 times when merchants planned into high-demand weeks (1.23-1.42 times across eight variants). It selected the right weeks, but inventory ordered to its lift differed from the true lift by 0.44-0.48. Its stated response to lasting regular-price changes was 4.7-6.5 times the known effect; only a measured-cause approach recovered it. On Dominick's, a noisy benchmark from 38 episodes estimated a pooled eight-week response of -0.33 (95% interval -0.68 to 0.01). Tuned world models and LightGBM stated about twice that measured response, and deeper price cuts lowered forecasts in only 29%-83% of item-week steps. Across 703 settings, check pass counts were not associated with forecast accuracy (Spearman 0.02, 95% interval -0.02 to 0.05); explanation correctness could be assessed only in generated data with known effects.

### Source excerpt

arXiv:2611.20811v1 Announce Type: cross Abstract: Forecasting models that read the price and promotion plan explain each forecast as a baseline, what would sell with nothing planned, plus the lift the plan adds. We test such explanations with eight checks, six with a pass rule fixed before any result, and find that passing them does not make an explanation right. In generated store data with known effects, a world model passes the checks of construction, stability, price direction and agreement with sales and with LightGBM, yet its lifts are 1.33 to 1.40 times the truth when merchants plan into high-demand weeks, and 1.23 to 1.42 times in eight variants of that world. It still picks the right weeks to promote, but stock ordered to its lift is off by 0.44 to 0.48 of the true lift. Its stated response to a lasting regular-price change is 4.7 to 6.5 times the true -0.5, and only a measured cause finds this. On Dominick's, 38 episodes in which a store's regular prices sat apart from its zone peers give a noisy benchmark: a pooled 8-week response of -0.33 (95% interval -0.68 to 0.01). For the final tuned world models of our companion paper on nine data sets, baseline and lift add up under each model's own plan, but a deeper price cut lowers the forecast in 29% to 83% of item-week steps on Dominick's (23% to 80% within prices seen in training), and the world model and LightGBM both state about twice the measured response. Across 703 configurations from the settings searches, checks passed were not related to forecast accuracy (Spearman 0.02, 95% interval -0.02 to 0.05); with no known answer for those versions, this concerns accuracy, while whether an explanation is right is measured only in the generated data.

[Next page](<https://devfeed.tech/topics/time-series.md?cursor=WyIyMDI2LTExLTE3VDA1OjAwOjAwKzAwOjAwIiwgIjdlMDljNjdhLWMwYzctNGRiMC1hM2I2LTMzZjFiMGZhYWJkMCJd>)