Research Intern: Affective Computing for Collective Intelligence - Honda Research Institute USA

Research Intern: Affective Computing for Collective Intelligence

Your application is being processed

Research Intern: Affective Computing for Collective Intelligence

Job Number: P25INT-50
Honda Research Institute USA (HRI-US) is seeking a research intern to work on an exciting machine learning challenge: understanding and modeling the objectives and dynamics that drive human team behavior. Current approaches focus on predicting team performance from observed interaction. This project aims to go a step further by modeling what teams are implicitly optimizing. The intern will develop computational models to draw on techniques such as inverse reinforcement learning, preference learning, or related ML methods, to uncover the latent objectives governing team dynamics such as information sharing, participation balance, and consensus formation. The project integrates ideas from multi-agent learning, collective intelligence, and behavioral psychology, and leverages real-world multimodal datasets of small group collaboration. The expected outcome is a research contribution suitable for submission to top-tier AI conferences (e.g., NeurIPS, ICML, ICLR), with potential impact on both human-AI interaction and multi-agent systems.
San Jose, CA

 

Key Responsibilities

 
  • Develop computational models (inverse reinforcement learning (IRL) or preference learning models) to infer latent reward functions from team interaction data.
  • Design and implement theory-informed reward components inspired by collective intelligence and behavioral psychology (such as information sharing, participation balance, consensus dynamics).
  • Process and model multimodal interaction datasets, including conversational transcripts, temporal interaction signals, and computer vision features.
  • Generate and evaluate conterfactual or alternative interaction trajectories for IRL training.
  • Conduct experiments on team perfomance prediction, early outcome inference, and generalization across datasets.
  • Analyze learned reward strctures to identify distinct team strategies and behavioral patterns.

 

Minimum Qualifications

 
  • Current PhD candidate in Computer Science, Electrical Engineering, Behavioral Psychology, or a related technical field.
  • Strong programming skills in Python and experiences with deep learning frameworks.
  • Solid understanding of machine learning fundamentals, including supervised learning and model evaluation.
  • Familiarity with sequence modeling (such as RNN, Transformers) or probabilistic modeling.

 

Bonus Qualifications​

  • Proven publication record with reinforcement learning, inverse reinforcement learning, imitation learning, or preference learning.
  • Background in multi-agent systems, game theory, or sequential decision-making.
  • Experience with multimodal or time-series data.
  • Familiarity with natural language processing or conversational modeling.
  • Interest in human behavior, social interaction, or computational social science.
  • Prior research experience or publications in machine learning or related areas.

 

Years of Work Experience Required   0
Desired Start Date  9/8/2026
Internship Duration  3 Months
Position Keywords  Machine Learning, Affective Computing, Human State Sensing, Multimodal Data Analysis, Collective Intelligence 

Alternate Way to Apply

Send an e-mail to careers@honda-ri.com with the following:
- Subject line including the job number(s) you are applying for 
- Recent CV 
- A cover letter highlighting relevant background (Optional)

Please, do not contact our office to inquiry about your application status.