Can We Build Large-Scale Models for Recommendation Systems?
In this talk, Zhang Pengtao, a Sina Weibo technical expert with a Ph.D. in computer applications, explores how the strong memory capabilities of NLP large language models inspire the design of independent memory mechanisms for recommendation systems, covering model concepts, HCNet & MemoNet, experimental results, and practical takeaways for enhancing recommendation model performance.
šØāš¼ Zhang Pengtao, Sina Weibo Technical Expert
Personal introduction: Ph.D. in Computer Applications from Peking University, author of multiple machineālearning papers, currently working on recommendation algorithms and user growth at Sina Weibo, focusing on personalized recommendation systems, large language models, user profiling, and data mining.
š„ Talk Title: Can We Build LargeāScale Models for Recommendation Systems
Talk Outline: The strong memory capability is one of the reasons why NLP large language models perform exceptionally well; this inspires us to construct an independent memory mechanism in recommendation models to store, learn, and recall arbitrary feature combinations, thereby improving recommendation performance. The main content includes:
1. Starting from NLP large models
2. HCNet & MemoNet
3. Experimental results
š Audience Gains:
1. What insights do NLP large models provide for recommendation model design?
2. How to enhance the memory capability of recommendation models?
3. Does strengthening recommendation model memory significantly affect model performance?
DataFunTalk
Dedicated to sharing and discussing big data and AI technology applications, aiming to empower a million data scientists. Regularly hosts live tech talks and curates articles on big data, recommendation/search algorithms, advertising algorithms, NLP, intelligent risk control, autonomous driving, and machine learning/deep learning.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.