SCHEDULE
August 28, 2024, 11:00AM–16:30PM (CEST), Barcelona, Spain.
Centre de Convencions Internacional de Barcelona
This will be a hybrid session. Zoom link will be provided later.
Opening
11:00-11:30
Introduction by organizers.
Moderator: Yilun Jin, Ph.D. Student at Hong Kong University of Science and Technology
Speaker: Dr. Zheng Li, Senior Applied Scientist at Amazon
Invited Speaker
11:30-12:10
Professor Huan Liu
Regents Professor, Arizona State University
Title: Social Media Mining: Embracing LLMs and Beyond
Abstract: Social media data differs from conventional data in many ways. It is big, noisy, linked, multimodal, and user generated. Through the lens of social media, unprecedented opportunities emerge for researchers in knowledge discovery and data mining. The prevalent usage of Large Language Models (LLMs) has added new challenges. We will use some of the challenges to illustrate the immediate need for (1) pondering what we should do to make LLMs more trustworthy and usable, and (2) contemplating a future beyond LLMs.
Invited Speaker
12:10-12:50
Professor Jingbo Shang
Associate Professor, University of California, San Diego
Title: Incubating Text Classifiers Following User Instruction with Nothing but LLM
Abstract: Automated knowledge extraction and discovery methods can address the diverse needs of different users. A fundamental open problem is how much user effort automated methods require to obtain useful knowledge. My research introduces a novel paradigm, extremely weak supervision (XWS), aimed at minimizing user effort. XWS involves only brief natural-language input from users to define tasks, such as a list of topics for classifying news articles, akin to the guidelines given to human annotators. In this talk, using text classification as an example, we will present a series of XWS methods. These methods encompass three primary approaches to pseudo-labeling: (1) mining-based X-Class, (2) generation-based Incubator, and (3) a hybrid approach, Text Grafting.
Lunch Break
12:50-14:00
Oral Presentations
14:00-16:00
- Team: Team_NVIDIA
Title: Winning Amazon KDD Cup'24
Time: 14:00 - 14:25 Slides
- Team: AML_LabCityU
Title: Second Place Overall Solution for Amazon KDD Cup 2024
Time: 14:25 - 14:50 Slides
- Team: shimmering_as_the_stars
Title: LLaSA: Large Language and E-Commerce Shopping Assistant
Time: 14:50 - 15:00 Video - Team: BMI_DLUT
Title: Enhancing User Behavior Alignment by Input-Level Model Cooperation and Model-Level Parameter Optimization
Time: 15:00 - 15:15 Slides - Team: ZJU_AI4H
Title: EC-Guide: A Comprehensive E-Commerce Guide for Instruction Tuning and Quantization by ZJU-AI4H
Time: 15:15 - 15:25 Video Slides - Team: Innova
Title: Fine-Tuning Large Language Models for Multitasking in Online Shopping Using Synthetic Data
Time: 15:25 - 15:40 Slides - Team: CM_RLLM
Title: Tailoring LLMs for Online Shopping: A Strategy for Multi-Task Learning and Domain-Specific Performance Enhancement
Time: 15:40 - 15:50 Video
ACCEPTED PAPERS
Winning Amazon KDD Cup'24
Track 1 Track 2 Track 3 Track 4 Track 5Authors: Chris Deotte, Ivan Sorokin, Ahmet Erdem, Benedikt Schifferer, Gilberto Titericz Jr, Simou Jegou,
Second Place Overall Solution for Amazon KDD Cup 2024
Track 1 Track 2 Track 3 Track 4 Track 5Authors: Pengyue Jia, Jingtong Gao, Xiaopeng Li, Zixuan Wang, Yiyao Jin, Xiangyu Zhao
LLaSA: Large Language and E-Commerce Shopping Assistant
Track 1 Track 4Authors: Shuo Zhang, Boci Peng, Xinping Zhao, Boren Hu, Yun Zhu, Yanjia Zeng, Xuming Hu
Tailoring LLMs for Online Shopping: A Strategy for Multi-Task Learning and Domain-Specific Performance Enhancement
Track 5Authors: Liu Yankai, Hao Yifan, Guo Ruipu, Cui Zhaojun
EC-Guide: A Comprehensive E-Commerce Guide for Instruction Tuning and Quantization by ZJU-AI4H
Track 2Authors: Zhaopeng Feng, Zijie Meng, Zuozhu Liu.
Enhancing User Behavior Alignment by Input-Level Model Cooperation and Model-Level Parameter Optimization
Track 3Authors: Yingyi Zhang, Zhipeng Li, Zhewei Zhi, Xianneng Li
Fine-Tuning Large Language Models for Multitasking in Online Shopping Using Synthetic Data
Authors: Fernando Sebastián Huerta, Carla Martín Monteso, Carlos de Leguina León, Leyre Sánchez Viñuela, Julián Rojo García, Rubén Sordo López, Roberto Lara Martin.More diverse more adaptive: Comprehensive Multi-task Learning for Improved LLM Domain Adaptation in E-commerce
Authors: Piao Tong, Pei Tang, Zhipeng Zhang, Jiaqi Li, Qiao Liu, Zufeng Wu.Optimizing Few-Shot Learning: From Static to Adaptive in Qwen2-7B
Authors: Wenhao Liu, Tianxing Bu, Erchen Yu, Dailin Li, Ding Ai, Zhenyi Lu, Haoran Luo.SUBMISSION GUIDELINES
The objective of this workshop is to discuss the winning submissions of the Amazon KDD Cup 2024: Multi-task Online Shopping Challenge for LLMs.
Submissions to the workshop are single-blind (author names and affiliations should be listed).
A team will have a guaranteed opportunity for an in-person oral/poster presentation if the team ranks in top-5 in any one of the 5 tracks.
Other submissions will be evaluated by a committee based on their novelty and insights.
The deadline for the submissions is August 2, 2024 August 4, 2024 (Anywhere on Earth time).
Accepted submissions will be notified latest by August 6, 2024.
Please note that the KDD Cup workshop will have no proceedings and the authors retain full rights to submit or post the paper at any other venue.
Link to the submission website: https://openreview.net/group?id=KDD.org/2024/Workshop/Amazon_KDD_Cup
\documentclass[sigconf, review]{acmart}.
Template guidelines are here: https://www.acm.org/publications/proceedings-template.
In addition, authors can provide an optional one page supplement at the end of their submitted paper (it needs to be in the same PDF file) focused on reproducibility. After the submission deadline, the names and order of authors cannot be changed.
Our dataset paper is available at ArXiv, and the Shopping MMLU dataset is available at github. Please cite our paper if you find our work helpful.
SUBMISSION DATES
Link to the submission website: https://openreview.net/group?id=KDD.org/2024/Workshop/Amazon_KDD_Cup
ORGANIZERS
DATA
The data and its license is available at github.
If you plan to use this dataset for your own research, please cite this paper.
@article{jin2024shopping,
title={Shopping MMLU: A Massive Multi-Task Online Shopping Benchmark for Large Language Models},
author={Jin, Yilun and Li, Zheng and Zhang, Chenwei and Cao, Tianyu and Gao, Yifan and Jayarao, Pratik and Li, Mao and Liu, Xin and Sarkhel, Ritesh and Tang, Xianfeng and others},
journal={arXiv preprint arXiv:2410.20745},
year={2024}
}