LMAct: A Benchmark for In-Context Imitation Learning with Long Multimodal Demonstrations

• URL: http://arxiv.org/abs/2412.01441v2
• Date: Mon, 03 Feb 2025 23:26:42 GMT
• Title: LMAct: A Benchmark for In-Context Imitation Learning with Long Multimodal Demonstrations
• Authors: Anian Ruoss, Fabio Pardo, Harris Chan, Bonnie Li, Volodymyr Mnih, Tim Genewein,
• Abstract summary: We present a benchmark to pressure-test today's decision-making models. We investigate whether models can learn numbers of expert demonstrations in their context.
• Score: 15.102187997350983
• License:
• Abstract: In this paper, we present a benchmark to pressure-test today's frontier models' multimodal decision-making capabilities in the very long-context regime (up to one million tokens) and investigate whether these models can learn from large numbers of expert demonstrations in their context. We evaluate the performance of Claude 3.5 Sonnet, Gemini 1.5 Flash, Gemini 1.5 Pro, Gemini 2.0 Flash Experimental, GPT-4o, o1-mini, o1-preview, and o1 as policies across a battery of simple interactive decision-making tasks: playing tic-tac-toe, chess, and Atari, navigating grid worlds, solving crosswords, and controlling a simulated cheetah. We study increasing amounts of expert demonstrations in the context $\unicode{x2013}$ from no demonstrations to 512 full episodes. Across our tasks, models rarely manage to fully reach expert performance, and often, presenting more demonstrations has little effect. Some models steadily improve with more demonstrations on a few tasks. We investigate the effect of encoding observations as text or images and the impact of chain-of-thought prompting. To help quantify the impact of other approaches and future innovations, we open source our benchmark that covers the zero-, few-, and many-shot regimes in a unified evaluation.

Related papers

• The Impact of Demonstrations on Multilingual In-Context Learning: A Multidimensional Analysis [23.757767581876063]
  In-context learning is a popular inference strategy where large language models solve a task using only a few labeled demonstrations. We show that the effectiveness of demonstrations varies significantly across models, tasks, and languages. We also find that strong instruction-following models including Llama 2-Chat, GPT-3.5, and GPT-4 are largely insensitive to the quality of demonstrations.
  arXiv  Detail & Related papers  (2024-02-20T12:53:31Z)
• Perception Test: A Diagnostic Benchmark for Multimodal Video Models [78.64546291816117]
  We propose a novel multimodal video benchmark to evaluate the perception and reasoning skills of pre-trained multimodal models. The Perception Test focuses on skills (Memory, Abstraction, Physics, Semantics) and types of reasoning (descriptive, explanatory, predictive, counterfactual) across video, audio, and text modalities. The benchmark probes pre-trained models for their transfer capabilities, in a zero-shot / few-shot or limited finetuning regime.
  arXiv  Detail & Related papers  (2023-05-23T07:54:37Z)
• Program Generation from Diverse Video Demonstrations [49.202289347899836]
  Generalising over multiple observations is a task that has historically presented difficulties for machines to grasp. We propose a model that can extract general rules from video demonstrations by simultaneously performing summarisation and translation.
  arXiv  Detail & Related papers  (2023-02-01T01:51:45Z)
• Out-of-Dynamics Imitation Learning from Multimodal Demonstrations [68.46458026983409]
  We study out-of-dynamics imitation learning (OOD-IL), which relaxes the assumption to that the demonstrator and the imitator have the same state spaces. OOD-IL enables imitation learning to utilize demonstrations from a wide range of demonstrators but introduces a new challenge. We develop a better transferability measurement to tackle this newly-emerged challenge.
  arXiv  Detail & Related papers  (2022-11-13T07:45:06Z)
• Robustness of Demonstration-based Learning Under Limited Data Scenario [54.912936555876826]
  Demonstration-based learning has shown great potential in stimulating pretrained language models' ability under limited data scenario. Why such demonstrations are beneficial for the learning process remains unclear since there is no explicit alignment between the demonstrations and the predictions. In this paper, we design pathological demonstrations by gradually removing intuitively useful information from the standard ones to take a deep dive of the robustness of demonstration-based sequence labeling.
  arXiv  Detail & Related papers  (2022-10-19T16:15:04Z)
• Rethinking the Role of Demonstrations: What Makes In-Context Learning Work? [112.72413411257662]
  Large language models (LMs) are able to in-context learn by conditioning on a few input-label pairs (demonstrations) and making predictions for new inputs. We show that ground truth demonstrations are in fact not required -- randomly replacing labels in the demonstrations barely hurts performance. We find that other aspects of the demonstrations are the key drivers of end task performance.
  arXiv  Detail & Related papers  (2022-02-25T17:25:19Z)
• Learning from Imperfect Demonstrations from Agents with Varying Dynamics [29.94164262533282]
  We develop a metric composed of a feasibility score and an optimality score to measure how useful a demonstration is for imitation learning. Our experiments on four environments in simulation and on a real robot show improved learned policies with higher expected return.
  arXiv  Detail & Related papers  (2021-03-10T07:39:38Z)
• Robust Maximum Entropy Behavior Cloning [15.713997170792842]
  Imitation learning (IL) algorithms use expert demonstrations to learn a specific task. Most of the existing approaches assume that all expert demonstrations are reliable and trustworthy, but what if there exist some adversarial demonstrations among the given data-set? We propose a novel general frame-work to directly generate a policy from demonstrations that autonomously detect the adversarial demonstrations and exclude them from the data set.
  arXiv  Detail & Related papers  (2021-01-04T22:08:46Z)
• Reinforcement Learning with Supervision from Noisy Demonstrations [38.00968774243178]
  We propose a novel framework to adaptively learn the policy by jointly interacting with the environment and exploiting the expert demonstrations. Experimental results in various environments with multiple popular reinforcement learning algorithms show that the proposed approach can learn robustly with noisy demonstrations.
  arXiv  Detail & Related papers  (2020-06-14T06:03:06Z)
• State-Only Imitation Learning for Dexterous Manipulation [63.03621861920732]
  In this paper, we explore state-only imitation learning. We train an inverse dynamics model and use it to predict actions for state-only demonstrations. Our method performs on par with state-action approaches and considerably outperforms RL alone.
  arXiv  Detail & Related papers  (2020-04-07T17:57:20Z)

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.