An official website of the United States government
Relevance to proposal: This project evaluates the generalizability of real and synthetic training datasets which can be used to train model-free techniques for multi-agent applications. We evaluate different methods of generating training corpora and machine learning techniques including Behavior Cloning and Generative Adversarial Imitation Learning. Our results indicate that the utility-guided selection of representative scenarios to generate synthetic data can have significant improvements on model performance. Paper abstract: Crowd simulation, the study of the movement of multiple agents in complex environments, presents a unique application domain for machine learning. One challenge in crowd simulation is to imitate the movement of expert agents in highly dense crowds. An imitation model could substitute an expert agent if the model behaves as good as the expert. This will bring many exciting applications. However, we believe no prior studies have considered the critical question of how training data and training methods affect imitators when these models are applied to novel scenarios. In this work, a general imitation model is represented by applying either the Behavior Cloning (BC) training method or a more sophisticated Generative Adversarial Imitation Learning (GAIL) method, on three typical types of data domains: standard benchmarks for evaluating crowd models, random sampling of state-action pairs, and egocentric scenarios that capture local interactions. Simulated results suggest that (i) simpler training methods are overall better than more complex training methods, (ii) training samples with diverse agent-agent and agent-obstacle interactions are beneficial for reducing collisions when the trained models are applied to new scenarios. We additionally evaluated our models in their ability to imitate real world crowd trajectories observed from surveillance videos. Our findings indicate that models trained on representative scenarios generalize to new, unseen situations observed in real human crowds.
more »
« less
This content will become publicly available on December 10, 2025
Learning to Cooperate with Humans using Generative Agents
Training agents that can coordinate zero-shot with humans is a key mission in multi-agent reinforcement learning (MARL). Current algorithms focus on training simulated human partner policies which are then used to train a Cooperator agent. The simulated human is produced either through behavior cloning over a dataset of human cooperation behavior, or by using MARL to create a population of simulated agents. However, these approaches often struggle to produce a Cooperator that can coordinate well with real humans, since the simulated humans fail to cover the diverse strategies and styles employed by people in the real world. We show \emph{learning a generative model of human partners} can effectively address this issue. Our model learns a latent variable representation of the human that can be regarded as encoding the human's unique strategy, intention, experience, or style. This generative model can be flexibly trained from any (human or neural policy) agent interaction data. By sampling from the latent space, we can use the generative model to produce different partners to train Cooperator agents. We evaluate our method -- \textbf{G}enerative \textbf{A}gent \textbf{M}odeling for \textbf{M}ulti-agent \textbf{A}daptation (GAMMA) -- on Overcooked, a challenging cooperative cooking game that has become a standard benchmark for zero-shot coordination. We conduct an evaluation with real human teammates, and the results show that GAMMA consistently improves performance, whether the generative model is trained on simulated populations or human datasets. Further, we propose a method for posterior sampling from the generative model that is biased towards the human data, enabling us to efficiently improve performance with only a small amount of expensive human interaction data.
more »
« less
- Award ID(s):
- 2212310
- PAR ID:
- 10634936
- Publisher / Repository:
- Neural Information Processing Systems 2024
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Simulators are being increasingly used to train agents before deploying them in real-world environments. While training in simulation provides a cost-effective way to learn, poorly modeled aspects of the simulator can lead to costly mistakes, or blind spots. While humans can help guide an agent towards identifying these error regions, humans themselves have blind spots and noise in execution. We study how learning about blind spots of both can be used to manage hand-off decisions when humans and agents jointly act in the real-world in which neither of them are trained or evaluated fully. The formulation assumes that agent blind spots result from representational limitations in the simulation world, which leads the agent to ignore important features that are relevant for acting in the open world. Our approach for blind spot discovery combines experiences collected in simulation with limited human demonstrations. The first step applies imitation learning to demonstration data to identify important features that the human is using but that the agent is missing. The second step uses noisy labels extracted from action mismatches between the agent and the human across simulation and demonstration data to train blind spot models. We show through experiments on two domains that our approach is able to learn a succinct representation that accurately captures blind spot regions and avoids dangerous errors in the real world through transfer of control between the agent and the human.more » « less
-
Most existing benchmarks for grounding language in interactive environments either lack realistic linguistic elements, or prove difficult to scale up due to substantial human involvement in the collection of data or feedback signals. We develop WebShop – a simulated e-commerce website environment with 1.18 million real-world products and 12,087 crowd-sourced text instructions. In this environment, an agent needs to navigate multiple types of webpages and issue diverse actions to find, customize, and purchase a product given an instruction. WebShop provides several challenges including understanding compositional instructions, query (re-)formulation, dealing with noisy text in webpages, and performing strategic exploration. We collect over 1,600 human trajectories to first validate the benchmark, then train and evaluate a diverse range of agents using reinforcement learning, imitation learning, and pre-trained image and language models. Our best model achieves a task success rate of 29%, which significantly outperforms rule heuristics but is far lower than expert human performance (59%). We also analyze agent and human trajectories and ablate various model components to provide insights for developing future agents with stronger language understanding and decision making abilities. Finally, we show our agent trained on WebShop exhibits non-trivial sim-to-real transfer when evaluated on amazon.com and ebay.com, indicating the potential value of our benchmark for developing practical web agents that can operate in the wild.more » « less
-
We present a novel data-driven simulation environment for modeling traffic in metropolitan street intersections. Using real-world tracking data collected over an extended period of time, we train trajectory forecasting models to learn agent interactions and environmental constraints that are difficult to capture conventionally. Trajectories of new agents are first coarsely generated by sampling from the spatial and temporal generative distributions, then refined using state-of-the-art trajectory forecasting models. The simulation can run either autonomously, or under explicit human control conditioned on the generative distributions. We present the experiments for a variety of model configurations. Under an iterative prediction scheme, the way-pointsupervised TrajNet++ model obtained 0.36 Final Displacement Error (FDE) in 20 FPS on an NVIDIA A100 GPU.more » « less
-
null (Ed.)Model-based reinforcement learning (RL), which finds an optimal policy using an empirical model, has long been recognized as one of the cornerstones of RL. It is especially suitable for multi-agent RL (MARL), as it naturally decouples the learning and the planning phases, and avoids the non-stationarity problem when all agents are improving their policies simultaneously using samples. Though intuitive and widely-used, the sample complexity of model-based MARL algorithms has been investigated relatively much less often. In this paper, we aim to address the fundamental open question about the sample complexity of model-based MARL. We study arguably the most basic MARL setting: two-player discounted zero-sum Markov games, given only access to a generative model of state transition. We show that model-based MARL achieves a near optimal sample complexity for finding the Nash equilibrium (NE) \emph{value} up to some additive error. We also show that this method is near-minimax optimal with a tight dependence on the horizon and the number of states. Our results justify the efficiency of this simple model-based approach in the multi-agent RL setting.more » « less
