The paper introduces Contrastive Noise Alignment (CNA), a training-time method for generative flow models that dynamically aligns Gaussian noise with data samples using a cross-modal InfoNCE objective. By modeling noise as an interacting particle system and regularizing with angular entropy and radial norm penalties, CNA reduces arbitrary data-noise couplings and flow curvature. Empirical results show that CNA improves generation quality, lowering FID by over 50% for few-step pixel-space generation compared to standard rectified flow and outperforming optimal transport baselines by at least 24%.
By Lennart Wittke, Vinicius Azevedo
The paper extends Mixed-Integer Differentiable Predictive Control (MI‑DPC) to handle multi‑modal discrete decisions and nonconvex polynomial dynamics in Underground Pumped Hydro Energy Storage Systems (UPHES). It introduces a neural policy that maps problem parameters to continuous setpoints and integer mode selections using a Gumbel‑Softmax layer, trained via self‑supervised differentiation of the finite‑horizon control objective. Key methodological advances include a parallel differentiable simulator, a Transformer encoder for long‑range temporal dependencies, and a Gumbel‑Softmax temperature annealing schedule, enabling the framework to achieve only 1.6% suboptimality compared to a mixed‑integer quadratic programming baseline while providing a five‑order‑of‑magnitude speedup in online scheduling.
By Honghui Zheng, J\'an Boldock\'y, Yury Dvorkin, J\'an Drgo\v{n}a
The report introduces TabPFN‑3.5, a new flagship tabular foundation model that outperforms its predecessor TabPFN‑3 and all existing baselines on a wide range of tabular tasks. It achieves state‑of‑the‑art performance on standard tabular prediction in TabArena and extends to practical scenarios such as non‑i.i.d. data, temporal or grouped splits, tables containing strings, text, images, high‑cardinality categorical features, and wide tables. Variants like TabPFN‑3.5‑Fast, TabPFN‑3.5‑Plus, and TabPFN‑3.5‑Thinking offer faster inference, expanded multimodal capabilities, and further speed improvements up to 12× faster than the previous Thinking mode.
By Benjamin J\"ager, Nick Erickson, L\'eo Grinsztajn, Felix Birkel, Klemens Fl\"oge, Oscar Key, K\"ur\c{s}at Kaya, Jonas K\"ubler, Ad\`ele Frankel, Tobias Schr\"oder, Anurag Garg, Jan Hendrik Metzen, David Salinas, Simon Bing, Kristina Collins, Tuana \c{C}elik, Vahid Balazadeh, Lydia Sidhoum, Tom\'as Pereda, Brendan Roof, Andrej Tschalzev, Siyuan Guo, Philipp Singer, Lennart Purucker, Jake Robertson, Marie Salmon, Philipp Jund, Jerry Chen, Diana Kriuchkova, Arthur Cahu, Eliott Kalfon, Adrian Hayler, Georg Grab, Vitor Monteiro, Lilly Wehrhahn, Dominik Safaric, Clara Cornu, Alan Arazi, Rylee Grace, Simone Alessi, Mihir Manium, Bernhard Sch\"olkopf, Yann LeCun, Madelon Hulsebos, Sauraj Gambhir, Noah Hollmann, Frank Hutter
The paper introduces MCLC‑NET, a multimodal continual learning framework for leaf counting that sequentially learns tasks using a memory buffer to retain key samples. It also presents MMLC, a new real‑world dataset containing RGB, depth, and thermal images across different crops and environmental conditions, organized in crop‑wise, time‑wise, and mixed orderings. Experiments show that MCLC‑NET outperforms existing methods on all three task orderings, achieving the lowest average mean squared errors.
By Ruchi Bhatt, Pratibha Kumari, Shreya Bansal, Vedant Agnihotri, Dwarikanath Mahapatra, Mukesh Saini
arXiv:2609.18084v1 Announce Type: cross
Abstract: Fine-tuning a Vision-Language-Action (VLA) model for a new deployment environment is expensive, yet most methods apply uniform-capacity adapters to e...
By Shahram Najam Syed, Arthur Jakobsson, Prayuj Sachdev, Jeffrey Ichnowski
TeleAntiFraud 2.0 is a monthly‑frozen, audio‑based benchmark for telecom fraud detection that incorporates newly observed scam patterns while preserving earlier test sets. It uses a Mixed‑Tree Anti‑Fraud Generation Pipeline to create profile‑grounded scenarios, expands them into mixed‑tree dialogues, and renders validated speech for 900 Chinese calls (600 fraud, 300 near‑domain non‑fraud) each month. Experiments show that classifiers perform well against unrelated negatives but drop significantly against near‑domain negatives, highlighting the need for near‑domain construction and collapse‑aware reporting in realistic evaluation settings.
By Huiyuan Liu, Zhiming Ma, Yanxing Liu, Shun Zhang, Qifan Wang, Di Liu, Yifan Wang, Yuyang Deng, Haoyang Meng, Yijin Zhou, Yuxi Zhao, Chengxian Hu, Peidong Wang, Peng Chen
The paper investigates why large vision‑language models sometimes misclassify harmful memes, attributing failures to either missing internal evidence or poor routing of evidence to the output. Using sparse autoencoders, role‑conditioned probes, and causal interventions on Gemma‑3 and Qwen3.5, the authors show that sparse readouts consistently outperform native predictions across six harmful content benchmarks, revealing a readout gap that is largely due to routing rather than representation. The study also demonstrates that calibration‑only routing recovers most of the performance gap and that the issue persists across languages and is not solely driven by OCR signals.
By Girish A. Koushik, Diptesh Kanojia, Helen Treharne
The paper introduces label‑free bias‑only test‑time reinforcement learning (TTRL), which uses majority‑vote pseudo‑labels as rewards and optimizes only about 100 K bias parameters while keeping the pretrained backbone frozen. On the MATH‑500 benchmark it achieves 76.67 % accuracy, slightly better than a labeled bias‑steering baseline, and improves performance on several vision‑language and audio reasoning tasks. The authors also show that the learned steering vectors transfer to 4,500 held‑out MATH problems and analyze why such a highly restricted adaptation works, linking majority‑vote reliability to rollout consensus and gradient energy in bias subspaces.
By Naveen Vakada, Mingyuan Li, Shaoxiong Ji
The paper introduces rMuscle, a real‑time Vision‑Language‑Action inference framework that mimics human muscle memory to accelerate robotic decision making. By exploiting repeated task similarity, rMuscle uses a dual‑phase cache: a Context Cache reuses visual‑token outputs and an Action Cache reuses neuron activation patterns, reducing computation and weight accesses. Experiments on RTX 4090 and Jetson Thor show 1.29–1.42× speedups on LIBERO, RoboTwin, and physical manipulation tasks while preserving success rates on real robots.
By Kaijun Zhou, Zhiyang Li, Le Chen, Jinyu Gu
This paper proposes a Retrieval-Augmented Generation (RAG) framework for scientific image quality assessment, designed to simultaneously address both the understanding track (SIQA-U) and the scoring t...
Artificial intelligence for plant disease analysis has advanced from task-specific classifiers to multi-modal models capable of jointly interpreting visual and textual information. However, practical...
Safety tuning pipelines judge only the final answer, which makes it difficult to distinguish robust refusal from two undesirable shortcuts: blanket refusal on benign requests and polished but unfaithf...
Intelligent systems that act in the world require image understanding that is both comprehensive and spatially grounded. Current vision-language models (VLMs) can generate fluent and detailed image ca...
Factory work is a promising early scenario for embodied AI: assigning repetitive manual jobs to robots has clear economic payoff, and a structured station keeps the jobs tractable for current policies...
Multi-modal talking avatar synthesis aims to generate realistic talking videos from a reference portrait and speech. Despite rapid progress in diffusion-based methods, existing approaches still strugg...
Self-driving laboratories can explore synthesis conditions autonomously, but their decision-making layer is typically a black-box optimizer, and the output is a set of optimized samples, with the meas...
Automatic speech recognition (ASR) systems exhibit unequal error rates across speaker groups, motivating interventions on their internal representations. We ask whether speaker-linked attributes that...
While Large Language Models excel in natural language processing, efficiently extending their capabilities to spoken input remains a significant challenge. Existing methods for building SpeechLLMs oft...
Vision-Language-Action models and World-Action Models have advanced language-conditioned robotic manipulation, yet often leave metric relations among actions, objects, and scene geometry implicit. Hum...
In Taiwanese Hokkien automatic speech recognition (ASR), prior studies often treat tone sandhi as a major challenge under the assumption that models fail to process implicit phonological variations. H...