Test-time Computing: from System-1 Thinking to System-2 Thinking
Ji, Yixin, Li, Juntao, Ye, Hai, Wu, Kaixin, Xu, Jia, Mo, Linjian, Zhang, Min
–arXiv.org Artificial Intelligence
The remarkable performance of the o1 model in complex reasoning demonstrates that test-time computing scaling can further unlock the model's potential, enabling powerful System-2 thinking. However, there is still a lack of comprehensive surveys for test-time computing scaling. We trace the concept of test-time computing back to System-1 models. In System-1 models, test-time computing addresses distribution shifts and improves robustness and generalization through parameter updating, input modification, representation editing, and output calibration. In System-2 models, it enhances the model's reasoning ability to solve complex problems through repeated sampling, self-correction, and tree search. We organize this survey according to the trend of System-1 to System-2 thinking, highlighting the key role of test-time computing in the transition from System-1 models to weak System-2 models, and then to strong System-2 models. We also point out a few possible future directions.
arXiv.org Artificial Intelligence
Jan-5-2025
- Country:
- Asia (1.00)
- North America > United States (0.46)
- Genre:
- Overview (1.00)
- Industry:
- Education (0.46)
- Technology: