Reinforcement learning (RL) has become a key component in modern large language models, yet the rollout stage remains the key bottleneck in RL training pipelines. Although Multi-Token Prediction (MTP) offers a natural solution to accelerate rollouts through speculative decoding, many studies have observed that MTP acceptance rates degrade significantly during RL training, leading to limited speedup performance.
Neural network pruning reduces model size by removing less important parameters while aiming to preserve predictive performance. Although the Lottery Ticket Hypothesis (LTH) shows that sparse subnetworks can match dense networks when trained from suitable initializations, its iterative pruning procedure requires multiple complete training cycles.
We propose a multi-agent collaborative framework built upon a lightweight Multimodal Large Language Model (MLLM), specifically designed for social intelligence reasoning. A key feature of our approach is that both the training and inference phases are augmented via knowledge distillation.
On-policy self-distillation (OPSD) provides dense, token-level supervision for reasoning models by aligning a model's own distribution with the distribution it produces under privileged context, typically a verified solution. However, we show that the learning signal drawn from this distributional gap concentrates on style tokens rather than task-bearing ones, as the hinted model tends to produce more direct, shorter outputs.
Deep search requires agents to answer complex questions through multi-step web search, browsing, evidence comparison, and synthesis. A central challenge is deciding how to search when several directions look plausible but only some will later lead to reliable evidence.
arXiv:2606. 10916v1 Announce Type: cross Abstract: This paper introduces range regularization for federated learning with linear systematic components to enhance statistical accuracy and induce cross-client regularity conducive to quantization, coding, and resource efficiency.
By Yiyuan She, Zhaojun Hu, Yifan Sun
arXiv:2606. 10209v1 Announce Type: new Abstract: Large language models deployed as autonomous agents for enterprise workflows face a key challenge: verbose tool responses from enterprise systems can cause context overflow, stale-state errors, and high inference cost.
By Abhilasha Lodha, Mahsa Pahlavikhah Varnosfaderani, Abir Chakraborty, Abhinav Mithal
arXiv:2602. 22352v2 Announce Type: replace-cross Abstract: With the continuous growth of neural network scales, low-precision quantization is widely used in edge accelerators.
By Yuhao Liu, Salim Ullah, Akash Kumar
arXiv:2606. 10532v1 Announce Type: new Abstract: Memory is essential for enabling large language model (LLM) agents to handle long-horizon reasoning tasks.
By Yunhan Jiang, Wenbin Duan, Shasha Guo, Liang Pang, Xiaoqian Sun, Huawei Shen
arXiv:2606. 10154v1 Announce Type: new Abstract: Quantized checkpoints are often screened first with quality metrics and only later, if at all, with direct safety tests.
By Sahil Kadadekar
arXiv:2606. 10613v1 Announce Type: cross Abstract: Diffusion-based Q-learning has emerged as a powerful paradigm for offline reinforcement learning, but its reliance on multi-step denoising makes both training and inference computationally expensive and brittle.
By Thanh Nguyen, Tri Ton, Hongbin Choe, Tung M. Luu, Chang D. Yoo
arXiv:2606. 09937v1 Announce Type: cross Abstract: We introduce RKSC (Reasoning-Aware KV Cache Sharing), a training-free inference framework that eliminates two structural redundancies in multi-branch LLM reasoning pipelines.
By Anirudh Sekar
arXiv:2606. 09922v1 Announce Type: cross Abstract: Bioelectrical signals are increasingly acquired at scales that challenge the bandwidth of brain-computer interfaces.
By Jiawen Zou, Bo Yan
arXiv:2606. 11033v1 Announce Type: cross Abstract: Recent efforts to extend large language models (LLMs) to speech inputs typically rely on cascaded ASR-LLM pipelines, end-to-end speech-language models, or bridge/distillation-based adaptation.
By Bo Cheng, Lei Shi, Zhanyu Ma, Yuan Wu, Jun Xu, Jiuchong Gao, Jinghua Hao, Renqing He
arXiv:2606. 09919v1 Announce Type: cross Abstract: Perceptual uncertainty is a central challenge for heterogeneous robot teams operating in unstructured outdoor environments, where no single viewpoint affords reliable scene understanding.
By Michal P. Podolinsky, Neel P. Bhatt, Pranay Samineni, Rohan Siva, Christian Ellis, Ufuk Topcu
arXiv:2606. 09916v1 Announce Type: cross Abstract: Multi-turn LLM agents fan short queries into long trajectories of tool calls, search results, and intermediate reasoning.
By Junjie Li, Jiong Lou, Jie Li
arXiv:2606. 10219v1 Announce Type: cross Abstract: AI efficiency at scale is becoming critical in finance as market data volumes surge across equities, ETFs, FX, options, and high-frequency trading streams.
By Henry Han, Diane Li
arXiv:2606. 10705v1 Announce Type: cross Abstract: Reinforcement learning promises to optimize sequential decisions in large-scale systems.
By Yavar Yeganeh, Mahsa Shekari, Nicla Frigerio, Daniele Pagano, Andrea Matta
arXiv:2504. 03118v2 Announce Type: replace-cross Abstract: Vision Transformers (ViTs) often need to be compressed for deployment on resource-constrained edge devices like drones and smart vehicles.
By Ziteng Wei, Qiang He, Bing Li, Feifei Chen, Hai Jin, Yun Yang
arXiv:2606. 10698v1 Announce Type: cross Abstract: In this paper, we use machine learning to discover a new seeding strategy for integration-by-parts reduction of Feynman integrals, which is a frequent bottleneck in state-of-the-art calculations in theoretical particle and gravitational-wave physics.
By Justin Berman, Francois Charton, Andres Luna, Matthias Wilhelm, Mao Zeng