KS

Kunyu Shi

cs.CVcs.AIcs.CLcs.LGeess.IV

On Valency

published · living versions

W_zbunfz7y·v1 · currentpublished

Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-work

with Wenhui Chen, Shiwen Cheng, Hao Dong, Chenda Duan +38

1 version

Preprints & journals

10 papers in the corpus · 2019–2026

Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-work2609.11977v1 · Wenhui Chen, Shiwen Cheng, Hao Dong et al.2026 · 0 citationsarXivon Valency

Business Arena: Benchmarking LLM Agents in a Realistic Marketplace2608.08621v1 · Yijun Pan, Yukun Lian, Kunyu Shi et al.2026 · 0 citationsarXiv

Act Wisely: Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Models2604.08545v1 · Shilin Yan, Jintao Tong, Hongwei Xue et al.2026 · 0 citationsarXiv

Beyond Scattered Acceptance: Fast and Coherent Inference for DLMs via Longest Stable Prefixes2603.05454v1 · Pengxiang Li, Joey Tsai, Hongwei Xue et al.2026 · 0 citationsarXiv

SwimBird: Eliciting Switchable Reasoning Mode in Hybrid Autoregressive MLLMs2602.06040v1 · Jintao Tong, Shilin Yan, Hongwei Xue et al.2026 · 0 citationsarXiv

Hybrid Reward Normalization for Process-supervised Non-verifiable Agentic Tasks2509.25598v1 · Peiran Xu, Zhuohao Li, Xiaoying Xing et al.2025 · 0 citationsarXiv

Enhancing Vision-Language Pre-training with Rich Supervisions2403.03346v2 · Yuan Gao, Kunyu Shi, Pengkai Zhu et al.2024 · 6 citationsarXiv

Non-autoregressive Sequence-to-Sequence Vision-Language Models2403.02249v2 · Kunyu Shi, Qi Dong, Luis Goncalves et al.2024 · 3 citationsarXiv

Musketeer: Joint Training for Multi-task Vision Language Model with Task Explanation Prompts2305.07019v2 · Zhaoyang Zhang, Yantao Shen, Kunyu Shi et al.2023 · 1 citationarXiv

Learning Instance Occlusion for Panoptic Segmentation1906.05896v4 · Justin Lazarow, Kwonjoon Lee, Kunyu Shi et al.2019 · 76 citationsarXiv

Career total: 10 works. 10 are in this corpus.

Profile built from the corpus for this byline.

Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.