regression

The paper introduces D^CF5, a diagnostic to predict regionwise gains in dynamic ensembling for regression tasks under distribution shift, validated with high correlation across datasets.

The article reports on performance regression in Grok 4.6 compared to 4.5, with lower pass rate and higher latency, affecting practical business tasks.

This paper tests whether perceived attributes of consumer technologies, rated by humans and frontier language models, predict ownership prevalence better than years-since-launch, finding modest improvements but limitations for short-term forecasts.

This paper proposes a new asymmetric robust bounded sparse smooth (aR) loss function for l1-norm penalized geometric twin support vector machine (aRSGTSVM) to handle classification and regression tasks with label and feature noise, achieving feature selection and robustness. Experiments on synthetic and UCI datasets plus China stock market index tracking demonstrate superiority.

This paper introduces TS2TabPFN, a framework that combines explicit feature extraction with the TabPFN 2.5 tabular foundation model for time series classification and extrinsic regression. Experiments show it outperforms state-of-the-art models in TSER and achieves competitive results in TSC.

This paper systematically compares tabular foundation models (TabPFN) with classical regression approaches across 85 soil spectroscopy tasks, finding that TabPFN combined with PLS-derived features achieves the best predictive performance from field-scale to global spectral libraries.

This paper studies the efficacy of various Graph Neural Network message-passing layers in regression contexts, finding that deep convolutional GNNs, particularly GEN, outperform attention-based GNNs.

The article evaluates the upgrade from Gemini 3.5 Flash to 3.6 Flash, noting aggregate benchmark gains but potential regressions in certain tasks, and recommends rigorous evaluation with predeclared failure gates before upgrading.

This paper proposes a two-stage online learning framework for detecting service-affecting failures in mobile core networks by modeling normal traffic dynamics and analyzing residuals, achieving improved precision-recall trade-off over static thresholds.

This paper presents a systematic comparison of two geospatial foundation models, TerraMind and THOR, developed under ESA's φ-lab, analyzing how architectural choices like patch size and decoder type affect performance across ten use cases in Earth observation tasks.

The paper proposes a lightweight method that reformulates regression-based INR training as a classification task by discretizing continuous targets into bins, enabling flexible distribution modeling for error-aware uncertainty estimation in scientific data compression.

Zer0Fit provides an MCP server that wraps Google's TabFM and TimesFM foundation models for zero-shot forecasting, classification, and regression tasks, running entirely locally.

OpenAI's model experiences a significant regression on the SimpleBench benchmark, indicating a drop in performance.

This paper investigates distributed sketching for OLS regression, where sketches are built from partitioned subsets rather than the whole dataset, reducing computational cost. The authors characterize the exact excess loss of the averaged estimator and show it matches that of whole-data sketching when subset covariance divergence is small.

Introduces an efficient Bayesian deep ensemble method for predictive regression that combines low-dimensional ensemble representation, closed-form Bayesian aggregation, and independent ensemble training to achieve calibrated uncertainty estimates with computational efficiency.

Linux 7.0 removed the PREEMPT_NONE kernel preemption mode, causing a PostgreSQL benchmark regression in specific configurations, but most users will not notice the change.

Newer Anthropic models like Opus 4.8 and Sonnet 5 are worse at using third-party editing tools (e.g., Pi's) compared to older models, likely because they were trained to use Claude Code's built-in edit tool via RL, causing them to invent extra fields in tool calls.

Newer Claude models (Opus 4.8 and Sonnet 5) exhibit worse tool-calling behavior by inventing extra fields in tool invocation arguments, causing validation failures, a regression compared to older models.

Third-party tests show that after returning, the Fable 5 model's performance dropped significantly on the BridgeBench benchmark. Debugging score fell from 86.2 to 25.9, Refactoring from 73.6 to 38.4, Hallucination from 75.9 to 61.7. It is speculated that the new safety guardrails caused many tasks to be handed off to Opus 4.8.

Proposes Geometry-aware R-Structured KAN (GRS-KAN), a hybrid neural architecture that integrates R-functions into KAN to encode geometric and logical constraints, achieving up to 67% RMSE reduction on regression benchmarks with discontinuities.