TSCA-Net proposes three complementary modules (Temporal-Spatial Clique Attention, Cross-Pedestrian Clique Potential, Adaptive KAN Grid Refinement) to improve pedestrian trajectory prediction in crowded environments, achieving state-of-the-art results on ETH/UCY and SDD benchmarks.

Two hybrid continual learning methods, RA-EWC and CGKD, are proposed to identify endangered Australian Aboriginal languages using limited data, outperforming fine-tuning and existing CL baselines on Warlpiri, Dalabon, and Dharawal.

A new paper reveals how large language models separate a character's belief from reality through a value slot and routing mechanism. The research shows two distinct mechanisms: a generic value slot for binding attributed values, and a router that selects which frame (belief or reality) to read from. The slot is filled either directly or via visibility-gated lookback, and the separation resides in routing subspaces rather than the slot itself. These findings hold across architectures and emerge between 3B and 7B parameters.

The paper introduces MAGE, a framework for studying component interactions in prompt optimization, discovering the Prompt Optimization Coupling Effect (POCE) where multiple stochastic signals interact to improve performance while amplifying variance. Key findings: failure-grounded reflection is crucial; MAGE outperforms GEPA on GSM8K-Hard; increasing candidate diversity amplifies POCE; POCE is headroom-dependent; in low-data settings, fixed prompts outperform reflective optimizers.

This work fine-tunes LLaMA 3 (8B) as an efficient drop-in reranker via supervised fine-tuning and 4-bit quantization, replacing cross-encoders in RAG pipelines. It achieves 14-21% improvement in answer relevancy, context precision, answer similarity, and answer correctness on a domain-specific QA benchmark while reducing inference overhead.

This paper proposes a text dataset distillation framework using Trajectory-Aware Knowledge Estimation (TAKE), which leverages influence functions and training trajectories to compute knowledge scores for each sample, and selects prototypes via optimal transport, compressing corpora to 0.1% of original size while preserving downstream task fidelity. Evaluated on text classification and natural language inference tasks.

This paper proposes a graph-based framework combining weak supervision with propagation graph analysis for detecting disinformation narratives in Telegram ecosystems. It aggregates semantically related claims into narrative-level clusters and models their diffusion across interconnected channels, enabling scalable detection of coordinated amplification.

A study evaluating state-of-the-art LLMs on Korean-Braille translation finds poor, unstable outputs and weak human agreement. Fine-tuning a small T5-small model yields large gains, revealing systematic limitations in current LLMs for accessibility-critical modalities.

This paper proposes G-SHARE, a framework that operationalizes the CNNP nine-step human-factor event diagnosis guideline into a multi-stage pipeline including evidence extraction, stepwise reasoning, and post-hoc consistency repair. Evaluated on real nuclear industry data, G-SHARE significantly outperforms one-shot prompting and machine learning baselines, demonstrating the value of structured reasoning and consistency enforcement for robust diagnosis.

This paper introduces CANDI-QA, a novel dataset for evaluating LLMs in specialized domains like medical diagnostics and financial advisory. It includes information assistance and applied inference questions. Over ten models are evaluated, and MTSS-Net, a lightweight neuro-symbolic framework, is proposed as a baseline. Findings highlight challenges in achieving contextual alignment in niche domains.

This paper shows that scaling can substantially narrow the performance gap between point-in-time language models and their unconstrained counterparts. The authors trained decoder-only transformers with up to 4 billion parameters on 1 trillion chronologically filtered tokens from FineWeb, creating monthly model checkpoints from 2013 to 2024. On reasoning and understanding benchmarks, these models approach the performance of leading open-weight models of comparable size (e.g., Gemma-3-4B and LLaMA-7B) trained on temporally unrestricted data. Instruction fine-tuning via LoRA further improves downstream usability. The complete pipeline is released for reproducibility.

This paper proposes a generalized distribution-free semi-supervised learning framework using risk rewrite, extending PNU learning to multiclass classification. It derives minimum achievable variance, showing lower variance in asymmetric loss scenarios, and introduces two practical SSL methods that match or outperform existing approaches.

A hybrid physics-probabilistic learning framework, BattVAE-GP, combines VAE and Gaussian processes to efficiently model lithium-ion battery degradation trajectories under varying charging rates, providing uncertainty quantification and reducing computational costs.

A matched-budget audit of six KV-cache compression methods reveals that query visibility significantly alters performance rankings. Under the query-agnostic protocol (compression before seeing questions), only KeyDiff consistently outperforms simple baselines, while popular methods like SnapKV can underperform naive strategies.

As large pre-trained models scale up, fine-tuning under limited memory becomes challenging. LoRA reduces learnable parameters via low-rank matrices, but activation storage remains a bottleneck. CARE-LoRA replaces full input activations with low-rank compressed activations from the LoRA branch, and computes a lightweight reconstruction matrix for backpropagation, substantially reducing memory while maintaining or improving performance.

A new book by Afonso S. Bandeira, Amit Singer, and Thomas Strohmer covers the mathematical foundations of data science, including high-dimensional phenomena, dimensionality reduction, regression, classification, deep learning, and more across 16 chapters.

Mirror Theory proposes that intelligent systems be evaluated by their capacity for coherent continuations under repeated reflection. Viable Path Entropy (VPE) operationalizes this as a finite-budget measure. Experiments on Qwen2.5 models show that increasing token budget expands verified reachability and VPE, with the smaller 1.5B model outperforming 3B at higher budgets, suggesting capability is not parameter count but accessible continuation capacity.

This paper proposes a qubit-efficient quantum framework for hyperdimensional computing decomposition that reduces representation cost from O(D) to O(log D) by introducing logarithmic hypervector and binding encodings, combined with a modified Dürr-Høyer search procedure, maintaining O(√N^F) search complexity while achieving up to 2000× qubit reduction.

Single-shot fringe projection profilometry (FPP) networks that regress depth directly can exploit a shape-prior shortcut, recovering depth from object boundaries rather than from fringe phase. PhiCalNet outputs a wrapped-phase representation (sinϕ, cosϕ) and maps it to depth through a fixed differentiable calibration layer, removing the shape-prior solution architecturally. On a photorealistic synthetic benchmark, PhiCalNet reduces object MAE 3.3x to 4.46 mm, with residuals confined to 0.103% of pixels at the wrap discontinuity, and a three-frame extension reaches 1.16 mm. Pixel-wise conformal uncertainty quantification, the first for FPP, localizes error at the same discontinuity.

Linear attention replaces softmax attention's growing KV cache with a fixed recurrent state, but this compression limits exact state tracking and long-context memory. This paper introduces Semidirect Fourier Delta Attention (SFDA), a phase-controlled generalization of Kimi Delta Attention that replaces real diagonal decay with block-rotational Fourier control. The main result is a constructive chunk-WY factorization, enabling exact affine chunk transfer, formal stability and complexity bounds, and a compact characterization of phase-plus-low-rank memory. Experiments show SFDA learns cyclic memory while the phase-disabled KDA baseline remains near chance.

OmniPM-Net is a novel fusion model combining graph neural network station forecasts with chemical transport model gridded forecasts using a Convolutional Conditional Neural Process framework. Evaluated on 1,618 stations across China in 2024, it matches the best GNN accuracy while reducing CAMS MAE by 30% and providing continuous spatial fields. It excels during dust storms and high-concentration events.

This paper presents GenAI Evaluation, a configuration-driven pipeline for large-scale evaluation of retail conversational systems. It processes production logs via normalization, sharding, asynchronous execution, and schema-constrained LLM scoring, evaluating helpfulness, truthfulness, clarity, tone alignment, and translation. Selective re-evaluation handles only invalid records; schema locking and versioned configs ensure auditability. The pipeline processes ~50,000 records daily and has evaluated over 2 million interactions. Validation on 12,980 human-labeled records achieved macro F1 0.93 and 89% translation accuracy.

This study investigates how the interaction graph structure in multi-agent language model systems affects consensus formation. Using a naming-game protocol, researchers analyzed convention formation in open-weight LM populations (1.1B-32B parameters). They found that homophilous threshold-similarity routing exacerbates fragmentation, while bridge-seeking routing can repair fragmentation when memory is available. In heterogeneous populations, threshold-similarity fails to produce consensus, while state-component and label-disagreement bridges recover consensus. In homogeneous populations, retained history generally promotes consensus, with Qwen2.5-32B reaching stable consensus in all retained-history settings.

The paper introduces an 'agent-ready website' design framework to enhance e-commerce platforms for AI agents. Experiments show that agent-ready websites improve strict success rates from 49.3% to 89.3%, reduce partial outcomes from 43 to 3, and lower average step count from 9.31 to 6.49.

This paper presents LP Mining with LP2Graph, a method that mines the structure of published LP and MILP formulations into a reproducible dataset and an induced taxonomy, applied to railway rescheduling.

This paper critiques the culture of optimization in AI, arguing that while optimization can measure improbability in generated text, it cannot distinguish between error and invention. Despite this limitation, optimization has assumed the authority to define legitimate language within half a decade, replacing traditional institutions.

A reproducible calibration-first reward audit framework is proposed for smart greenhouse reinforcement learning control, decomposing scalar reward into conditional temperature, CO2, humidity, and actuation terms, validated on GreenLight-Gym and Autonomous Greenhouse Challenge data.

GRID is a grammar-constrained decoding engine that uses LALR(1) parser states as mask keys to ensure syntactically valid SQL output with role-based access control, provable guarantees, constant per-token cost, and a hash-chained audit trail. On Spider, constrained decoding boosts execution accuracy by +13 points at 0.5B, and a checker-guided repair lifts a 7B model to 94.5% executable.

This study combines ontology-amplified distillation and contextuality auditing for building and governing tenant-owned language models in regulated financial institutions. The distillation experiment shows a Qwen3.6-27B student grounds 36/40 Vietnamese financial tasks, matching GPT-5, but is underpowered to establish equivalence. A contextuality audit pilot finds zero residual contextuality, suggesting direct influence and construct coupling are more useful signals. The evidence does not support deployability, safety, or superiority.

This paper surveys in-context reinforcement learning (ICRL) under non-stationarity, where pretrained decision models infer latent task rules and improve behavior from interaction context without parameter updates. In changing environments, accumulated context can become stale or misleading, requiring the policy to infer both the current decision rule and which past evidence is still valid. The literature is organized around three questions: what changes, how the change unfolds, and how observable the change is, linking ICRL to meta-RL, decision sequence modeling, retrieval-augmented RL, and related approaches.