25 papers across AI, ML, NLP, and CV from the last 24 hours.
Today's batch reveals two convergent pressures reshaping AI research: efficiency at scale and safety in coordination. A cluster of papers attacks inference and training overhead from complementary angles — adaptive speculative decoding that varies proposal length based on compression state (SpecKV), joint parameter compression with fine-tuning rather than sequential application (JACTUS), and layer-wise training monitoring for low-bit transformers. The message is clear: we are past the phase of building bigger models and into the phase of making them tractable.
The second theme is multi-agent dynamics. Papers on LLM orchestration traces, multi-agent reinforcement learning through temporal interaction graphs, and text-to-SQL agents with flexible database exploration all treat agents not as isolated tools but as components in a larger system that must coordinate, delegate, and recover from errors. This is the infrastructure layer for the agent era.
The standout paper identifies misalignment contagion — the phenomenon where misaligned behavior spreads between language models during multi-turn interactions. This reframes alignment as a networked safety problem rather than a per-model property, which is both more realistic and more alarming given the current pace of multi-agent deployment.
Notably absent today: computer vision submissions were rate-limited out, but among the papers that got through, cs.CV is thin with only one entry. The stat.ML side is unusually strong, with DNN universality proofs via the Lindeberg exchange principle, Bayesian online learning framed through the analogy of an autonomic nervous system, and multiple works on robust optimization under heavy-tailed noise.
Collectively, the batch suggests the field is maturing from model construction to system governance — we are no longer asking what models can do, but how to deploy, monitor, and align them when they must operate together.
Mohamad Khajezade, Fatemeh H. Fard, Mohamed Sami Shehata · 2026-05-04
Cross-language code clone detection (X-CCD) is challenging because semantically equivalent programs written in different languages often share little surface similarity. Although large language models (LLMs) have shown promise for semantic clone detection, their use as black-box systems raises concerns about cost, reproducibility, privacy, and unreliable output formatting. In particular, compact o
Xinhui Xiong, Bin Gao, P. -A. Absil · 2026-05-04
Retraction-free approaches offer attractive low-cost alternatives to Riemannian methods on the Stiefel manifold, but they are often first-order, which may limit the efficiency under high-accuracy requirements. To this end, we propose a second-order method landing on the Stiefel manifold without invoking retractions, which is proved to enjoy local quadratic (or superlinear for its inexact variant)
Nidhi Vakil, Hadi Amiri · 2026-05-04
Structure aware graph generation aims to generate graphs that satisfy given topological properties. It has applications in domains such as drug discovery, social network modeling, and knowledge graph construction. Unlike existing methods that only provide coarse control over graph properties, we introduce a novel conditional variational autoencoder for fine-grained structural control in graph gene
Pehuén Moure, Niclas Pokel, Bilal Bounajma · 2026-05-04
Automatic speech recognition (ASR) systems remain brittle on dysarthric and other atypical speech. Recent audio-language models raise the possibility of improving performance by conditioning on additional clinical context at inference time, but it is unclear whether these models can make use of such information. We introduce a benchmark built on the Speech Accessibility Project (SAP) dataset that
Shikhar Shukla · 2026-05-04
Speculative decoding accelerates large language model (LLM) inference by using a small draft model to propose candidate tokens that a larger target model verifies. A critical hyperparameter in this process is the speculation lengthγ, which determines how many tokens the draft model proposes per step. Nearly all existing systems use a fixedγ (typically~4), yet empirical evidence suggests that
Bogdan Oancea · 2026-05-04
Ensuring the coherence of regional socio-economic statistics is a central task for national statistical institutes. Traditional validation tools, such as range edits, ratio checks, or univariate outlier detection, are effective for identifying extreme values in individual series but are less suited for detecting unusual combinations of indicators in high-dimensional settings. This paper proposes a
Haizhou Wen, Elham Kiyani, Gang Li · 2026-05-04
Composite materials exhibit strongly hierarchical and anisotropic properties governed by coupled mechanisms spanning constituents, plies, laminates, structures, and manufacturing history. This intrinsic complexity makes predictive modeling of composites expensive, because repeated experiments and high-fidelity simulations are needed to cover large design spaces of material, structure, and manufact
Lingxiao Kong, Cong Yang, Oya Deniz Beyan · 2026-05-04
Despite significant advances in Reinforcement Learning (RL), model performance remains highly sensitive to algorithm and hyperparameter configurations, while generalization gaps across environments complicate real-world deployment. Although prior work has studied RL generalization, the relative contribution of specific configurations to the generalization gap has not been quantitatively decomposed
Quang Hieu Pham, Yang He, Ping Nie · 2026-05-04
Text-to-SQL over large analytical databases requires navigating complex schemas, resolving ambiguous queries, and grounding decisions in actual data. Most current systems follow a fixed pipeline where schema elements are retrieved once upfront and the database is only revisited for post-hoc repair, limiting recovery from early mistakes. We present FlexSQL, a text-to-SQL agent whose core design pri
Chenchen Zhang · 2026-05-04
As large language model (LLM) agents evolve from isolated tool users into coordinated teams, reinforcement learning (RL) must optimize not only individual actions but also how work is spawned, delegated, communicated, aggregated, and stopped. This paper studies RL for LLM-based multi-agent systems through orchestration traces: temporal interaction graphs whose events include sub-agent spawning, de
Mario Rodríguez Béjar, B. Romera-Paredes, Jose L. Hernández-Ramos · 2026-05-04
Modern fuzzers increasingly use Large Language Models (LLMs) to generate structured inputs, but LLM-driven fuzzing is sensitive to prompt initialization and sampling variance, which can reduce exploration efficiency and lead to redundant inputs. We present FunFuzz, a multi-island evolutionary fuzzing framework that runs several isolated searches in parallel and periodically migrates high-value can
Dominik Macko · 2026-05-04
SemEval-2026 Task 10 is focused on conspiracy detection. Specifically, the goal is to detect whether a Reddit comment expresses a conspiracy belief. Our submitted mdok-style system utilizes data augmentation and self-training (to cope with a rather small amount of training data) to finetune the Qwen3-32B model for a binary text-classification task. The submitted system is very competitive, ranking
Tanush Yadav, Mohammadreza Salehi, Jae Sung Park · 2026-05-04
Videos are unique in their ability to capture actions which transcend multiple frames. Accordingly, for many years action recognition was the quintessential task for video understanding. Unfortunately, due to a lack of sufficiently diverse and challenging data, modern vision-language models (VLMs) are no longer evaluated on their action recognition capabilities. To revitalize action recognition in
Verena Jasmin Hallitschke, Carsten Eickhoff, Philipp Berens · 2026-05-04
Vision-language models hold considerable promise for ophthalmology, but their development depends on large-scale, high-quality image-text datasets that remain scarce. We present PubMed-Ophtha, a hierarchical dataset of 102,023 ophthalmological image-caption pairs extracted from 15,842 open-access articles in PubMed Central. Unlike existing datasets, figures are extracted directly from article PDFs
Durba Bhattacharya, Sucharita Roy, Sourabh Bhattacharya · 2026-05-04
This chapter introduces the Bayesian reflex -- an analogy with the autonomic nervous system -- as a unifying framework for online learning in AI. Bayesian online algorithms automatically maintain equilibrium in dynamic environments via three mechanisms: belief maintenance through probabilistic representations, sequential updating via Bayes' theorem, and uncertainty-driven action balancing explorat
Marcos Matabuena, Mateo Cámara · 2026-05-04
Across many scientific disciplines, multiple observations are collected from the same experimental units, and in modern datasets these observations often arise as non-Euclidean random objects. In such settings, the incorporation of random effects is a critical modeling step for efficient estimation and personalized prediction. Although mixed-effects models are well established for scalar outcomes
Yuxi Cai, Lan Li, Feiqing Huang · 2026-05-04
The proliferation of large-scale and structurally complex data has spurred the integration of machine learning methods into statistical modeling. Recurrent neural networks (RNNs), a foundational class of models for time-dependent data, can be viewed as nonlinear extensions of classical autoregressive moving average models. Despite their flexibility and empirical success in machine learning, RNNs o
This digest is generated automatically from arXiv submissions. Not affiliated with arXiv or Cornell University.