Good Papers

Showing papers from Cisco Research Show all papers

89%Must read
?Must readVote to see the score
NeurIPS 2026NortheasternU GeorgiaCisco ResearchQuantization

ActQuant: Sub-4-bit Action-Guided Quantization for Vision-Language-Action Models

ActQuant uses action-guided mixed-precision quantization to compress vision-language-action models below 4 bits, retaining 95% task performance at 3 bpw and enabling edge deployment via native C/C++ kernels.

Arash Akbari, Arman Akbari, Masih Eskandar, Qitao Tan and 10 more

Atlanta Poster Session 3, Thu, Dec 10, 10:00 AM–1:00 PM, Hall C1 · Published 2026 · ▲ 2 on Hugging Face · Code ★ 19

– ReadersNo votes yet
16/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 16 of 20 reviewers recommend it
lenient 5/5
medium 8/10
strict 3/5
83%Must read
?Must readVote to see the score
NeurIPS 2026U California, Santa BarbaraU Southern CaliforniaCiscoCisco ResearchRL for LLMs

Survive or Collapse: The Asymmetric Roles of Data Gating and Reward Grounding in Self-Play RL

Self-play RL stability depends mainly on a strict data gate over proposer tasks, not reward design; ground-truth access accelerates collapse via a self-consistent attractor.

Sophia Xiao Pu, Zhaotian Weng, Chengzhi Liu, Jayanth Srinivasa and 3 more

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet
13/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 13 of 20 reviewers recommend it
lenient 4/5
medium 7/10
strict 2/5
88%Must read
?Must readVote to see the score
NeurIPS 2026U California San DiegoU California, San DiegoCiscoCisco ResearchRL for LLMs

TIER: Trajectory-Invariant Execution Rewards for Multi-Step Tool Composition

TIER derives dense tool-use rewards from execution and schemas rather than reference paths, enabling over 90% accuracy on multi-step composition where trajectory supervision fails.

Anay Kulkarni, Chia En Lu, Dheeraj Mekala, Jayanth Srinivasa and 2 more

Atlanta Poster Session 4, Thu, Dec 10, 4:30 PM–7:30 PM, Hall C1 · Published 2026

– ReadersNo votes yet
15/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 15 of 20 reviewers recommend it
lenient 4/5
medium 10/10
strict 1/5