Yoel Drori
Research Subjects
Research Interests
Yoel Drori has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Daryl Chang
YouTube, Google, Mountain View, CA, USA
1 shared publication - Ed H. Chi
Google DeepMind, Mountain View, CA, USA
1 shared publication - Jiaxi Tang
Google DeepMind, USA
1 shared publication - Justin Gilmer
Google DeepMind, Mountain View, CA, USA
1 shared publication - Li Wei
YouTube, Google, Mountain View, CA, USA
1 shared publication - Lichan Hong
Google, Mountain View, CA, USA
1 shared publication
Latest Publications
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 4882-4893
Improving Training Stability for Multitask Ranking Models in Recommender Systems
Recommender systems play an important role in many content platforms. While most recommendation research is dedicated to designing better models to improve user experience, we found that research on stabilizing the training for such models is severely under-explored. As recommendation models become larger and more sophisticated, they are more susceptible to training instability issues, i.e., loss divergence, which can make the model unusable, waste significant resources and block model developments. In this paper, we share our findings and best practices we learned for improving the training stability of a real-world multitask ranking model for YouTube recommendations. We show some properties of the model that lead to unstable training and conjecture on the causes. Furthermore, based on our observations of training dynamics near the point of training instability, we hypothesize why existing solutions would fail, and propose a new algorithm to mitigate the limitations of existing solutions. Our experiments on YouTube production dataset show the proposed algorithm can significantly improve training stability while not compromising convergence, comparing with several commonly used baseline methods.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 4882-4893
Open AccessImproving Training Stability for Multitask Ranking Models in Recommender Systems
Recommender systems play an important role in many content platforms. While most recommendation research is dedicated to designing better models to improve user experience, we found that research on stabilizing the training for such models is severely under-explored. As recommendation models become larger and more sophisticated, they are more susceptible to training instability issues, i.e., loss divergence, which can make the model unusable, waste significant resources and block model developments. In this paper, we share our findings and best practices we learned for improving the training stability of a real-world multitask ranking model for YouTube recommendations. We show some properties of the model that lead to unstable training and conjecture on the causes. Furthermore, based on our observations of training dynamics near the point of training instability, we hypothesize why existing solutions would fail, and propose a new algorithm to mitigate the limitations of existing solutions. Our experiments on YouTube production dataset show the proposed algorithm can significantly improve training stability while not compromising convergence, comparing with several commonly used baseline methods.