TUNEAHEAD is a lightweight framework that predicts fine-tuning performance using meta-feature vectors from dataset descriptors and short probe runs. It outperforms baselines like Early-Stop Extrapolation and ProxyLM, achieving an RMSE of 1.47 percentage points and 95.1% of predictions within ±3 percentage points of true scores on 370 held-out runs.
A confusion-aware difficulty score is introduced within the Transfer Teacher framework to improve model interpretability and data efficiency. Evaluations on CIFAR-10 show that confusion-aware curriculum ordering outperforms random ordering by up to 8.7% at 20% data, demonstrating consistent data-efficiency gains. However, curriculum or anti-curriculum ordering does not improve accuracy over standard training at full data, indicating that scoring function improvements alone are insufficient to overcome curriculum learning failure modes.
SMAA-Fair extends Stochastic Multicriteria Acceptability Analysis by reweighting rankings based on group fairness. It incorporates fairness metrics like Statistical Parity, rKL, and nDKL, adjusting acceptability indices to better represent protected groups while maintaining robustness to preference uncertainty.
The paper proposes an unsupervised framework to recover latent domains and signals from corrupted observations by discovering data symmetries. It models observations as linear measurements of signals from a latent random field and uses a shallow group-convolutional network with stationarity and locality constraints to learn latent symmetry actions and filters, enabling recovery from unstructured data.
QueryMarket introduces OVBAL, an online variance-based active learning framework that estimates each data point's marginal utility using a D-optimality criterion with exponential forgetting. OVBAL selects samples based on utility and price, operating under rolling budget constraints and adapting to concept drift, showing improved error-cost trade-offs in solar power forecasting tasks.
The paper introduces 'No-Free-Fairness' theorems that prove three fundamental limits in learning systems. These include inherent fairness-cost trade-offs, unavoidable subgroup disparity in finite samples, and model expressivity constraints that prevent fairness regardless of data. The results show fairness is constrained by problem structure, data limits, and model capacity, not just biased data.
This paper proposes a meta-classification method for one-class classification models by representing them as normality rankings and using ranking correlation and nearest neighbor metrics. The approach achieves high accuracy in classifying models based on training datasets, algorithms, and hyperparameters, and works even when datasets share the same class. The method effectively classifies datasets by treating multiple samples as a single input, offering a unified solution for OCC models, datasets, and rankings.
McWC introduces a model that separately captures cyclicity, trend, and inter-channel correlations in long-term time series forecasting. It uses multi-layer cyclicity construction, wavelet decomposition, and a multi-layer perceptron to extract and fuse high- and low-frequency information, while decoupling intra-channel autocorrelations via frequency-domain loss. Experiments on six real-world datasets show McWC achieves state-of-the-art performance with high computational efficiency.
BLITZ introduces a two-stage regression method for nonparametric conditional independence testing. It first removes broad smooth dependencies using polynomial regression, then applies shallow tree regressions to residualize nonlinear features, enabling accurate and fast testing with improved null calibration compared to existing methods.
McWC introduces a model that separately captures cyclicity, trend, and inter-channel correlations in long-term time series forecasting. It uses multi-layer cyclicity construction, wavelet decomposition, and a multi-layer perceptron to extract and fuse high- and low-frequency information, while decoupling intra-channel autocorrelations via frequency-domain loss. Experiments on six real-world datasets show McWC achieves state-of-the-art performance with high computational efficiency.
A study of 14,727 security and privacy prompts from 3.2M real-world user-LLM conversations identifies nine categories of S&P questions. Thematic analysis and response testing show commercial LLMs outperform open models, with GPT 5.5 providing good responses on 98% of prompts versus Llama 4 at 47%, though some commercial models produce inconsistent responses across runs.
A study evaluated several AI systems on ten research-level mathematics problems created by prominent mathematicians. The results include AI-generated solutions, human solutions, and referee reports, offering a detailed assessment of AI performance in solving advanced mathematical problems.
Language models of GPT-2 size cannot independently discover zero during testing, regardless of pretraining. However, performance improves significantly with training on tens to hundreds of zero examples, and language pretraining reduces required examples by about 50%.
As enterprise agent tool catalogs scale from 10 to 110 agents, routing accuracy drops 16--23 percentage points on under-specified requests. An oracle analysis identifies retrieval and confusion gaps, with embedding-based shortlisting recovering +10--11pp F1. A human-annotated study of 1,435 utterances confirms real-world recovery of +10--17pp despite lower absolute performance.
A new framework uses prompt perturbation to identify and filter structurally inconsistent pairwise comparisons in large language model evaluations. By incorporating graph-level consistency checks before ranking aggregation, the method reduces cyclic preferences and improves the reliability of LLM rankings.
We present a framework for evaluating agentic skills by constructing realistic tasks and assessing skill utility through task execution. Applied to 500 real-world skills, it generates 1,000 tasks and scoring rubrics, evaluating 19 agent-model configurations across proprietary and open-source models. Results show significant variation in instruction adherence and performance gains, with skills substantially altering model behavior compared to no-skill setups.
A study finds bilingual fine-tuning enhances automatic speech recognition in low-resource languages when language identification is accurate. Including a language identification token at inference improves ASR performance when identification accuracy is low, especially in diverse language pairs across different families and writing systems.
NNN decoding selects documents as a joint set that jointly reconstructs the query embedding via a sparse non-negative linear combination. It strictly extends dense retrieval by handling queries that dense retrieval fails on, especially in corpora with correlated documents, and achieves superior performance through end-to-end training of embeddings.
Glimmer 1 is a 10,000-parameter language model trained on 500K tokens from FineWeb-Edu. It features a 512-token context window, a standard Llama architecture with 16 hidden dimensions, 2 layers, 4 attention heads, and 1 KV head using GQA, and is available on Hugging Face.
A measurement study finds that 26 semantic post-hoc operators do not improve held-out accuracy over Best-of-N in frozen small code models. While two operators—expression-layer recovery and adaptive consensus early-stop—offer benefits in compute efficiency or program recovery, none outperform BoN in accuracy. The results highlight systemic limitations in error detection and coverage, suggesting that model harnesses and error coverage must be improved before post-hoc reasoning is considered.