We explore using a moderately sized large language model (GPT-J 6B parameters) to create a plan for a simulated robot to achieve 30 classes of goals in ScienceWorld, a text game simulator for elementary science experiments. Previously published empirical work claimed that large language models (LLMs) are a poor fit (Wang et al., 2022) compared to reinforcement learning. Using the Markov assumption (a single previous step), the LLM outperforms the reinforcement learning-based approach by a factor of 1.4. When we fill the LLM's input buffer with as many prior steps as possible, improvement rises to 3.5x. Even when training on only 6.5% of the training data, we observe a 2.2x improvement over the reinforcement-learning-based approach. Our experiments show that performance varies widely across the 30 classes of actions, indicating that averaging over tasks can hide significant performance issues. In work contemporaneous with ours, Lin et al. (2023) demonstrated a two-part approach (SwiftSage) that uses a small LLM (T5-large) complemented by OpenAI's massive LLMs to achieve outstanding results in ScienceWorld. Our 6-B parameter, single-stage GPT-J matches the performance of SwiftSage's two-stage architecture when it incorporates GPT-3.5 turbo which has 29-times more parameters than GPT-J.
Neural Networks Remember More: The Power of Parameter Isolation and Combination
Aberdeen Institute of Data Science and Artificial Intelligence(阿伯丁数据科学与人工智能学院);South China Normal University(华南师范大学);School of Artificial Intelligence(人工智能学院)
Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
AGI Institute, School of Computer Science, Shanghai Jiao Tong University(AGI研究院,计算机科学学院,上海交通大学);Tongyi Lab, Alibaba Group(通义实验室,阿里巴巴集团);Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering, Shanghai Jiao Tong University(上海市教育委员会智能交互与认知工程重点实验室,上海交通大学)
When Modalities Remember: Continual Learning for Multimodal Knowledge Graphs
School of Computer Science, Peking University(北京大学计算机科学学院);School of Computer Science, Zhejiang University(浙江大学计算机科学学院);School of Cyber Science and Engineering, Wuhan University(武汉大学网络科学与工程学院);School of Information Science and Technology, Tibet University(西藏大学信息科学与技术学院)