aiqbench analysis
AffectOmni addresses a critical gap in multimodal large language models by enforcing verifiable attention to human cues like micro-expressions rather than relying on logical shortcuts. The framework utilizes specific reinforcement learning rewards and pixel-level grounding via SAM3 to create an externally auditable interface for affective reasoning tasks. This advancement improves traceability and performance on emotion recognition benchmarks, offering a more robust path toward reliable social AI applications.
AffectOmni: RL-Verifiable People-Centric Grounded Affective Reasoning for Social and Art-Related Scenes
arXiv:2608.26193v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) achieve strong performance on VQA and scene understanding, yet affective reasoning remains vulnerable to shortcut behavior. Models may predict correct answers while neglecting people-centric cues such as micro expressions and body language, which weakens traceability and external verification. Prior reinforcement learning approaches mainly reward context or logical coherence without explicitly enforcing attention to human evidence. In addition, LLM as a Judge scoring often suffers from score clustering, which reduces reward discriminability. We propose AffectOmni, a GRPO trained framework for verifiable affective reasoning. AffectOmni introduces People Focus and Temporal Order rewards to encourage people-centric evidence selection and temporally structured reasoning, and it adopts within-group comparative scoring to produce more stable and discriminative reward signals. For verification, a Thinking Summarizer converts free form rationales into executable evidence instructions, which are grounded into pixel level evidence regions via SAM3 to provide an externally auditable interface outside the training loop. Experiments on IntentBench, Daily Omni, and WorldSense show consistent improvements over open source 7B scale baselines, including gains of 4.66% on emotion recognition and +14.29% on temporally sensitive tasks. Code is available at https://github.com/eliot127825-rgb/AffectOmni_nobody.
Abstract:Multimodal large language models (MLLMs) achieve strong performance on VQA and scene understanding, yet affective reasoning remains vulnerable to shortcut behavior. Models may predict correct answers while neglecting people-centric cues such as micro expressions and body language, which weakens traceability and external verification. Prior reinforcement learning approaches mainly reward context or logical coherence without explicitly enforcing attention to human evidence. In addition, LLM as a Judge scoring often suffers from score clustering, which reduces reward discriminability. We propose AffectOmni, a GRPO trained framework for verifiable affective reasoning. AffectOmni introduces People Focus and Temporal Order rewards to encourage people-centric evidence selection and temporally structured reasoning, and it adopts within-group comparative scoring to produce more stable and discriminative reward signals. For verification, a Thinking Summarizer converts free form rationales into executable evidence instructions, which are grounded into pixel level evidence regions via SAM3 to provide an externally auditable interface outside the training loop. Experiments on IntentBench, Daily Omni, and WorldSense show consistent improvements over open source 7B scale baselines, including gains of 4.66% on emotion recognition and +14.29% on temporally sensitive tasks. Code is available at this https URL.
Artificial Intelligence (cs.AI)
arXiv:2608.26193 [cs.AI]
(or arXiv:2608.26193v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.26193 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 281 words
Related coverage
Game Design Creativity evals
LessWrong · Oct 12, 2026
Hint-based CoT faithfulness evals still mostly work on Claude
LessWrong · Jul 30, 2026
State of the Field: AI for Epistemics and Coordination
LessWrong · Oct 2, 2026
Microsoft’s AI chief says superintelligence is near, but won’t take your job
The Verge · Jun 8, 2026