# OmniSapiens: A Foundation Model for Social Behavior Processing via Heterogeneity-Aware Relative Policy Optimization

arXiv: https://arxiv.org/abs/2602.10635 · HTML: https://mlsanity.com/paper/2602.10635/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 45 of its 78 references that are in the map.
- First posted on arXiv 11 Feb 2026.

## Builds on (most cited first, 30 of 45)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [DAPO: An Open-Source LLM Reinforcement Learning System at Scale](https://mlsanity.com/paper/2503.14476/) (2025)
- [Trust Region Policy Optimization](https://mlsanity.com/paper/1502.05477/) (2015)
- [Gemma 3 Technical Report](https://mlsanity.com/paper/2503.19786/) (2025)
- [Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics](https://mlsanity.com/paper/1705.07115/) (2017)
- [Gradient Surgery for Multi-Task Learning](https://mlsanity.com/paper/2001.06782/) (2020)
- [Deep Reinforcement Learning that Matters](https://mlsanity.com/paper/1709.06560/) (2017)
- [Qwen2.5-Omni Technical Report](https://mlsanity.com/paper/2503.20215/) (2025)
- [MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversations](https://mlsanity.com/paper/1810.02508/) (2018)
- [Convex Optimization: Algorithms and Complexity](https://mlsanity.com/paper/1405.4980/) (2014)
- [Towards Robust Interpretability with Self-Explaining Neural Networks](https://mlsanity.com/paper/1806.07538/) (2018)
- [Qwen3-Omni Technical Report](https://mlsanity.com/paper/2509.17765/) (2025)
- [From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AI](https://mlsanity.com/paper/2201.08164/) (2022)
- [Distral: Robust Multitask Reinforcement Learning](https://mlsanity.com/paper/1707.04175/) (2017)
- [MOSI: Multimodal Corpus of Sentiment Intensity and Subjectivity Analysis in Online Opinion Videos](https://mlsanity.com/paper/1606.06259/) (2016)
- [A Survey of Reinforcement Learning for Large Reasoning Models](https://mlsanity.com/paper/2509.08827/) (2025)
- [Towards Multimodal Sarcasm Detection (An _Obviously_ Perfect Paper)](https://mlsanity.com/paper/1906.01815/) (2019)
- [GPG: A Simple and Strong Reinforcement Learning Baseline for Model Reasoning](https://mlsanity.com/paper/2504.02546/) (2025)
- [UR-FUNNY: A Multimodal Language Dataset for Understanding Humor](https://mlsanity.com/paper/1904.06618/) (2019)
- [What Do We Want From Explainable Artificial Intelligence (XAI)? -- A Stakeholder Perspective on XAI and a Conceptual Model Guiding Interdisciplinary XAI Research](https://mlsanity.com/paper/2102.07817/) (2021)
- [Foundations and Trends in Multimodal Machine Learning: Principles, Challenges, and Open Questions](https://mlsanity.com/paper/2209.03430/) (2022)
- [Do Androids Laugh at Electric Sheep? Humor "Understanding" Benchmarks from The New Yorker Caption Contest](https://mlsanity.com/paper/2209.06293/) (2022)
- [UniTS: A Unified Multi-Task Time Series Model](https://mlsanity.com/paper/2403.00131/) (2024)
- [GRPO-LEAD: A Difficulty-Aware Reinforcement Learning Approach for Concise Mathematical Reasoning in Language Models](https://mlsanity.com/paper/2504.09696/) (2025)
- [Composable Deep Reinforcement Learning for Robotic Manipulation](https://mlsanity.com/paper/1803.06773/) (2018)
- [AVEC 2016 - Depression, Mood, and Emotion Recognition Workshop and Challenge](https://mlsanity.com/paper/1605.01600/) (2016)

## Built on by (most cited first, 1 of 1)

- [Reinforcement Learning with Comparative Evidence for Social Intelligence](https://mlsanity.com/paper/2610.04072/) (2026)

## Abstract (arXiv)

Socially intelligent AI systems must reason across diverse human behavioral tasks and generalize to new social contexts. However, behavioral data is inherently heterogeneous, comprising diverse modalities and prediction targets that produce uneven training signals across samples, creating imbalanced learning dynamics that challenge existing AI models. To address this, we develop Omnisapiens-7B 2.0, a foundation model for social behavior processing that explicitly addresses learning from heterogeneous behavioral data. This is enabled through Heterogeneity-Aware Relative Policy Optimization, a new RL method that rebalances learning signals across samples by approximating each sample's contribution to the policy update and using these estimates to drive geometrically centered, inertially smoothed advantage modulation for stable training. Omnisapiens-7B 2.0 achieves the best and most consistent performance across 10 behavioral tasks, while also attaining the best performance on all five held-out benchmarks, with gains of up to +12.02% and +9.37% respectively. Furthermore, it demonstrates more consistent and interpretable reasoning traces, supporting reliable real-world behavioral applications. Our model is available at https://github.com/MIT-MI/human_behavior_atlas.