Stay updated with the latest research and technology news
arXiv:2510.22876v2 Announce Type: replace Abstract: Speculative decoding must produce outputs distribution identical to standard autoregressive generation-this output equivalence is not an optimization target but the...
arXiv:2510.22984v2 Announce Type: replace Abstract: Many scientific and geometric problems exhibit general linear symmetries, yet most equivariant neural networks are built for compact groups or...
arXiv:2510.23487v2 Announce Type: replace Abstract: This paper studies standard controller architectures for agentic AI and derives automata-theoretic models of their interaction behavior via trace semantics...
arXiv:2510.23912v4 Announce Type: replace Abstract: We theoretically investigate whether the Query, Key, Value weight triplet can be reduced in encoder-only and decoder-only transformers. Under mild...
arXiv:2510.24941v2 Announce Type: replace Abstract: Large language models can generate long chain-of-thought (CoT) reasoning, but it remains unclear whether the verbalized steps reflect the models'...
arXiv:2510.25128v2 Announce Type: replace Abstract: The technique of data augmentation (DA) is often used in machine learning for regularization purposes to better generalize under i.i.d....
arXiv:2510.25471v2 Announce Type: replace Abstract: Instrumental goals such as resource acquisition, power-seeking, and self-preservation are key to contemporary AI alignment research, yet the phenomenon's ontology...
arXiv:2510.25497v3 Announce Type: replace Abstract: Neurosymbolic AI is growing in popularity thanks to its ability to combine neural perception and symbolic reasoning in end-to-end trainable...
arXiv:2510.25801v3 Announce Type: replace Abstract: Reinforcement learning (RL) with verifiable rewards has recently catalyzed a wave of "MLLM-r1" approaches that bring RL to vision language...
arXiv:2510.26104v2 Announce Type: replace Abstract: In recommendation systems, scaling up feature-interaction modules (e.g., Wukong, RankMixer) or user-behavior sequence modules (e.g., LONGER) has achieved notable success....
arXiv:2510.26374v3 Announce Type: replace Abstract: Reinforcement finetuning (RFT) is a key technique for aligning Large Language Models (LLMs) with human preferences and enhancing reasoning, yet...
arXiv:2510.26412v2 Announce Type: replace Abstract: Recent advances in text-to-video generation have achieved impressive performance on short clips, yet evaluating long-form generation under complex textual inputs...
arXiv:2510.26555v2 Announce Type: replace Abstract: With the rapid advancement of information technology, the complexity of applications continues to increase, and the cybersecurity challenges we face...
arXiv:2510.26852v2 Announce Type: replace Abstract: Current evaluation for Large Language Model (LLM) code agents predominantly focus on generating functional code in single-turn scenarios, which fails...
arXiv:2510.27410v3 Announce Type: replace Abstract: A fundamental bottleneck in human-AI collaboration is the ``intention expression gap," the difficulty for humans to effectively convey complex, high-dimensional...
arXiv:2510.27647v3 Announce Type: replace Abstract: Collaborative perception improves task performance by expanding the perception range through information sharing among agents. . Immutable heterogeneity poses a...
arXiv:2511.00066v3 Announce Type: replace Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a practical route to improve large language model reasoning, and Group Relative...
arXiv:2511.00067v2 Announce Type: replace Abstract: The objective of domain generalization (DG) is to enable models to be robust against domain shift. DG is crucial for...
arXiv:2511.00655v2 Announce Type: replace Abstract: Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, yet its scalability is limited by...
arXiv:2511.01706v2 Announce Type: replace Abstract: Natural Language Explanations (NLEs) describe how Large Language Models (LLMs) make decisions by drawing on external Context Knowledge (CK) and...
arXiv:2511.02230v3 Announce Type: replace Abstract: KV cache management is essential for efficient LLM inference. To maximize utilization, existing inference engines evict finished requests' KV cache...
arXiv:2511.02767v2 Announce Type: replace Abstract: The alignment of representations from different modalities has recently been shown to provide insights on the structural similarities and downstream...
arXiv:2511.04993v2 Announce Type: replace Abstract: While the auto-bidding literature predominantly considers independent bidding, we investigate the coordination problem among multiple auto-bidders in online advertising platforms....
arXiv:2511.05005v2 Announce Type: replace Abstract: This work presents MAC-Flow, a simple yet expressive framework for multi-agent coordination. We argue that requirements of effective coordination are...