School of Engineering Department of Computer Science and Engineering 122 Reasoning with Large Foundation Models Supervisor: SONG Yangqiu / CSE Student: LAU Ching Ming Samuel / COGBM Course: UROP 1100, Spring Since the launch of ChatGPT, chatbots have demonstrated remarkable performance in different industries (Vaswani et al., 2017; Brown et al., 2020). In E-commerce, chatbots are developed such to answer user queries about products and recommend products to users. However, product descriptions and attributes can quickly become outdated, leading to errors of Large Language Model. Therefore, to ensure the relevance and accuracy of product information, there is a need to continuously update the knowledge of chatbots in this dynamic E-commerce landscape. In this research, we focus on tackling the challenge of updating the knowledge of LLM through knowledge editing in the E-commerce domain. First, we propose a new benchmark knowledge editing dataset for Ecommerce. Second, comprehensive experiments will be done on the proposed benchmark by using current knowledge editing methods. Lastly, we will explore a new knowledge editing method specifically for Ecommerce. Reasoning with Large Foundation Models Supervisor: SONG Yangqiu / CSE Student: LIU Jiayu / COMP WANG Hanwen / COMP Course: UROP 1000, Summer UROP 1000, Summer In the contemporary information age, the sheer volume of data is overwhelming, with vast amounts of content generated daily on the Internet. However, the reliability of these online statements remains a significant concern, profoundly impacting society. While it is feasible to manually annotate and verify the authenticity of online content, this process is both time-consuming and costly due to the immense quantity and rapid pace of information generation. Consequently, developing an automated system to verify the validity of online claims is both advantageous and essential, ensuring that individuals can fully utilize online information. Therefore, based on AVeriTeC, our team proposes a standardized methodology to assess the validity of claims using pertinent information. Reasoning with Large Foundation Models Supervisor: SONG Yangqiu / CSE Student: WONG Chun Ka / ISD Course: UROP 1000, Summer This paper reflects on my learning experience from Stanford’s CS 224N course on Natural Language Processing (NLP) with Deep Learning12. The course covered key topics such as Word Vectors, Neural Networks, Sequence Models, Attention Mechanisms, Machine Translation, and Sentiment Analysis3. Despite only engaging with the theoretical aspects, I gained valuable insights into NLP’s potential to revolutionize human-computer interaction. My initial fascination with NLP was sparked by the release of ChatGPT, highlighting the transformative power of AI in understanding and generating human language. This reflection underscores the importance of hands-on experience and a deeper understanding of the underlying principles to innovate and advance in the field of NLP.
RkJQdWJsaXNoZXIy NDk5Njg=