Read AI papers faster
with ChatPaper

EverMind AI · EverMind

As large language models advance, AI agents are moving beyond isolated, domain-specific tasks toward long-horizon, cross-domain workflows. This transition exposes two challenges:…

Haoyu Zhao, Zihao Zhang, Xudong Wang, Jiaxi Gu et al. · National University of Singapore

Executable programs offer explicit control over how images and videos are constructed, but generating runnable code is only the beginning of visual creation. A program can execute…

Haojian Huang, Zexi Li, Junhao Guo, Yehang Zhang et al. · Knowin AI

General-purpose robots must infer what a new task requires and translate that understanding into appropriate physical action. In-context learning (ICL) for robots supports this…

Yuntai Bao, Qinfeng Li, Guoqing Jiang, Liwei Chen et al. · Zhejiang University

*Reinforcement learning (RL)* can induce substantial reasoning capabilities in large language models (LLMs), but how much of this capability transfers across model scales, and how…

Yanlin Li, Mingyang Hao, Shengqiong Wu, Hao Fei et al. · National University of Singapore

General-purpose agents can plan, reason, and act over long horizons, yet their production capabilities remain fragmented across text, images, audio, video, documents, 3D assets,…

Zhen Wang, Changpeng Wang, Zhe Liu, Zhangyang Qi et al. · Zhejiang University

Recent vision-language models (VLMs) have advanced vision-and-language navigation (VLN), enabling models to predict navigation actions from visual observations and language…

Yang Xiao, Vidhyasaharan Sethu, Eun-Jung Holden, Ting Dang · The University of Melbourne

Spoken conversational systems must recover information from prior interactions (i.e., memory), yet relevant information in speech extends beyond what was said to who said it, how…

Xirui Li, Peng Shi, Mingwen Dong, Sheng Zhang et al. · Amazon Web Services

A 3D scene reconstructed from a single image is most useful when represented not as a rendering or a fixed 3D output, but as an explicit scene program whose execution yields a…

Renping Zhou, Zanlin Ni, Zihao Fan, Guohao Fu et al. · Tsinghua-LeapLab

World action models (WAMs) predict the future alongside actions during training. Due to the heavy computation cost of video denoising, whether the future must still be generated…

Hui Ren, Lei Fan, Henry Pao, Han Guo et al. · Amazon Science

Answering questions about long videos often requires connecting events involving the same objects across hours or days. Chronological descriptions and text-derived entities can…

Xuehai Bai, Zhenchen Tang, Yang Shi, Dianyi Wang et al.

Visual reward models are essential for evaluating and improving visual generation models, yet existing approaches typically map task conditions and candidate outputs directly to…

Xingyu Zhu, Pu, Yi, Ziheng Cheng et al. · ByteDance Seed

Chunked KV-cache compression reduces the memory and attention costs of long-context inference by compressing windows of consecutive tokens into fewer cache entries at a fixed…