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

• URL: http://arxiv.org/abs/2602.10635v1
• Date: Wed, 11 Feb 2026 08:35:59 GMT
• Title: OmniSapiens: A Foundation Model for Social Behavior Processing via Heterogeneity-Aware Relative Policy Optimization
• Authors: Keane Ong, Sabri Boughorbel, Luwei Xiao, Chanakya Ekbote, Wei Dai, Ao Qu, Jingyao Wu, Rui Mao, Ehsan Hoque, Erik Cambria, Gianmarco Mengaldo, Paul Pu Liang,
• Abstract summary: We introduce Heterogeneity-Aware Relative Policy Optimization (HARPO), an RL method that balances leaning across heterogeneous tasks and samples.<n>Using HARPO, we develop and release Omnisapiens-7B 2.0, a foundation model for social behavior processing.<n>Relative to existing behavioral foundation models, Omnisapiens-7B 2.0 achieves the strongest performance across behavioral tasks.
• Score: 50.11607985532808
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: To develop socially intelligent AI, existing approaches typically model human behavioral dimensions (e.g., affective, cognitive, or social attributes) in isolation. Although useful, task-specific modeling often increases training costs and limits generalization across behavioral settings. Recent reasoning RL methods facilitate training a single unified model across multiple behavioral tasks, but do not explicitly address learning across different heterogeneous behavioral data. To address this gap, we introduce Heterogeneity-Aware Relative Policy Optimization (HARPO), an RL method that balances leaning across heterogeneous tasks and samples. This is achieved by modulating advantages to ensure that no single task or sample carries disproportionate influence during policy optimization. Using HARPO, we develop and release Omnisapiens-7B 2.0, a foundation model for social behavior processing. Relative to existing behavioral foundation models, Omnisapiens-7B 2.0 achieves the strongest performance across behavioral tasks, with gains of up to +16.85% and +9.37% on multitask and held-out settings respectively, while producing more explicit and robust reasoning traces. We also validate HARPO against recent RL methods, where it achieves the most consistently strong performance across behavioral tasks.

Related papers

• Toward Enhancing Representation Learning in Federated Multi-Task Settings [13.483094713610322]
  Federated multi-task learning (FMTL) seeks to collaboratively train customized models for users with different tasks.<n>We propose Muscle loss, a contrastive learning objective that simultaneously aligns representations from all participating models.<n>We develop FedMuscle, a practical and communication-efficient FMTL algorithm that naturally handles both model and task.
  arXiv  Detail & Related papers  (2026-02-02T04:39:36Z)
• HumanLLM: Towards Personalized Understanding and Simulation of Human Nature [72.55730315685837]
  HumanLLM is a foundation model designed for personalized understanding and simulation of individuals.<n>We first construct the Cognitive Genome, a large-scale corpus curated from real-world user data on platforms like Reddit, Twitter, Blogger, and Amazon.<n>We then formulate diverse learning tasks and perform supervised fine-tuning to empower the model to predict a wide range of individualized human behaviors, thoughts, and experiences.
  arXiv  Detail & Related papers  (2026-01-22T09:27:27Z)
• Human Behavior Atlas: Benchmarking Unified Psychological and Social Behavior Understanding [50.34821397821815]
  Human Behavior Atlas comprises over 100,000 samples spanning text, audio, and visual modalities.<n>Our unification efforts can reduce redundancy and cost, enable training to scale efficiently across tasks, and enhance generalization of behavioral features across domains.
  arXiv  Detail & Related papers  (2025-10-06T15:16:45Z)
• Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Models [79.2162092822111]
  We systematically evaluate reinforcement learning (RL) and control-based methods on a suite of navigation tasks.<n>We employ a latent dynamics model using the Joint Embedding Predictive Architecture (JEPA) and employ it for planning.<n>Our results show that model-free RL benefits most from large amounts of high-quality data, whereas model-based planning generalizes better to unseen layouts.
  arXiv  Detail & Related papers  (2025-02-20T18:39:41Z)
• AlignDiff: Aligning Diverse Human Preferences via Behavior-Customisable Diffusion Model [69.12623428463573]
  AlignDiff is a novel framework to quantify human preferences, covering abstractness, and guide diffusion planning. It can accurately match user-customized behaviors and efficiently switch from one to another. We demonstrate its superior performance on preference matching, switching, and covering compared to other baselines.
  arXiv  Detail & Related papers  (2023-10-03T13:53:08Z)
• Multi-Timescale Modeling of Human Behavior [0.18199355648379031]
  We propose an LSTM network architecture that processes behavioral information at multiple timescales to predict future behavior. We evaluate our architecture on data collected in an urban search and rescue scenario simulated in a virtual Minecraft-based testbed.
  arXiv  Detail & Related papers  (2022-11-16T15:58:57Z)
• Task-Feature Collaborative Learning with Application to Personalized Attribute Prediction [166.87111665908333]
  We propose a novel multi-task learning method called Task-Feature Collaborative Learning (TFCL) Specifically, we first propose a base model with a heterogeneous block-diagonal structure regularizer to leverage the collaborative grouping of features and tasks. As a practical extension, we extend the base model by allowing overlapping features and differentiating the hard tasks.
  arXiv  Detail & Related papers  (2020-04-29T02:32:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.