Application of Pretrained Large Language Models in Embodied Artificial Intelligence

Abstract

A feature of tasks in embodied artificial intelligence is that a query to an intelligent agent is formulated in natural language. As a result, natural language processing methods have to be used to transform the query into a format convenient for generating an appropriate action plan. There are two basic approaches to the solution of this problem. One is based on specialized models trained with particular instances of instructions translated into agent-executable format. The other approach relies on the ability of large language models trained with a large amount of unlabeled data to store common sense knowledge. As a result, such models can be used to generate an agent’s action plan in natural language without preliminary learning. This paper provides a detailed review of models based on the second approach as applied to embodied artificial intelligence tasks.

A novel approach to voice of customer extraction using GPT-3.5 Turbo: linking advanced NLP and Lean Six Sigma 4.0

“…LLMs are deep neural network models trained on vast amounts of information, including books, code, articles, and websites, to grasp the underlying patterns and relationships in the language they are trained on [48]. Consequently, these models can generate coherent content, such as linguistically accurate sentences and paragraphs that resemble human language or structurally sound code snippets [49]. LLMs have numerous applications, including language translation, summarization, and question answering, and hold potential across various elds, provided the training data offers suitable input [50].…”

R-DDQN: Optimizing Algorithmic Trading Strategies Using a Reward Network in a Double DQN

“…In recent years, research based on RLHF (reinforcement learning with human feedback) has led to groundbreaking advancements in various fields. For instance, the remarkable progress achieved in the field of artificial intelligence, particularly with large language models [13,16,17], owes much to expert demonstrations or feedback used to assist model training. This approach enables models to cope with complex and dynamic environments.…”

Psychedelics and Artificial Intelligence: Integrating AI-Powered Analysis in Phenomenological Mental Health Studies

“…The convergence of Artificial Intelligence, Linguistics, and Cognitive Science as part of the field of Natural Language Processing (Lenci & Padó, 2022), the arrival of the Transformer architecture, based on attention mechanisms to capture contextual dependencies in sequential data (Vaswani et al, 2017), as well as the use of Large Language Models (LLMs) in contexts such as Embodied Artificial Intelligence (Kovalev & Panov, 2022), are examples of the last advances in the use of computational models specialized in human language (Hamilton et al, 2022).…”