Aviary: training language agents on challenging scientific tasks
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912172924207104 |
|---|---|
| author | Narayanan, Siddharth Braza, James D. Griffiths, Ryan-Rhys Ponnapati, Manu Bou, Albert Laurent, Jon Kabeli, Ori Wellawatte, Geemi Cox, Sam Rodriques, Samuel G. White, Andrew D. |
| author_facet | Narayanan, Siddharth Braza, James D. Griffiths, Ryan-Rhys Ponnapati, Manu Bou, Albert Laurent, Jon Kabeli, Ori Wellawatte, Geemi Cox, Sam Rodriques, Samuel G. White, Andrew D. |
| contents | Solving complex real-world tasks requires cycles of actions and observations. This is particularly true in science, where tasks require many cycles of analysis, tool use, and experimentation. Language agents are promising for automating intellectual tasks in science because they can interact with tools via natural language or code. Yet their flexibility creates conceptual and practical challenges for software implementations, since agents may comprise non-standard components such as internal reasoning, planning, tool usage, as well as the inherent stochasticity of temperature-sampled language models. Here, we introduce Aviary, an extensible gymnasium for language agents. We formalize agents as policies solving language-grounded partially observable Markov decision processes, which we term language decision processes. We then implement five environments, including three challenging scientific environments: (1) manipulating DNA constructs for molecular cloning, (2) answering research questions by accessing scientific literature, and (3) engineering protein stability. These environments were selected for their focus on multi-step reasoning and their relevance to contemporary biology research. Finally, with online training and scaling inference-time compute, we show that language agents backed by open-source, non-frontier LLMs can match and exceed both frontier LLM agents and human experts on multiple tasks at up to 100x lower inference cost. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_21154 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Aviary: training language agents on challenging scientific tasks Narayanan, Siddharth Braza, James D. Griffiths, Ryan-Rhys Ponnapati, Manu Bou, Albert Laurent, Jon Kabeli, Ori Wellawatte, Geemi Cox, Sam Rodriques, Samuel G. White, Andrew D. Artificial Intelligence Computation and Language Machine Learning Solving complex real-world tasks requires cycles of actions and observations. This is particularly true in science, where tasks require many cycles of analysis, tool use, and experimentation. Language agents are promising for automating intellectual tasks in science because they can interact with tools via natural language or code. Yet their flexibility creates conceptual and practical challenges for software implementations, since agents may comprise non-standard components such as internal reasoning, planning, tool usage, as well as the inherent stochasticity of temperature-sampled language models. Here, we introduce Aviary, an extensible gymnasium for language agents. We formalize agents as policies solving language-grounded partially observable Markov decision processes, which we term language decision processes. We then implement five environments, including three challenging scientific environments: (1) manipulating DNA constructs for molecular cloning, (2) answering research questions by accessing scientific literature, and (3) engineering protein stability. These environments were selected for their focus on multi-step reasoning and their relevance to contemporary biology research. Finally, with online training and scaling inference-time compute, we show that language agents backed by open-source, non-frontier LLMs can match and exceed both frontier LLM agents and human experts on multiple tasks at up to 100x lower inference cost. |
| title | Aviary: training language agents on challenging scientific tasks |
| topic | Artificial Intelligence Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2412.21154 |