Research and Applications of Foundation Models for Data Mining and Affective Computing (RAFDA)

7 - 10 May Taipei, Taiwan
PAKDD 2024 Workshop

Exciting News for RAFDA 2024

Exciting News: We are delighted to confirm that Springer will be the publisher for the main track of PAKDD 2024. Additionally, we are pleased to announce that the proceedings of our RAFDA workshop will also be published by Springer.
As promised, the proceedings of the RAFDA workshop have been published by Springer in the Lecture Notes in Artificial Intelligence (LNAI) series, and we have 4-week free access from now on: .

RAFDA Starts In:

Proceedings


PSpringer-LogoThe proceedings of the RAFDA workshop will be published by Springer (Lecture Notes in Artificial Intelligence (LNAI) series.
As promised, the proceedings of the RAFDA workshop have been published by Springer in the Lecture Notes in Artificial Intelligence (LNAI) series, and we have 4-week free access from now on: .

About RAFDA

This workshop, entitled Research and Applications of Foundation Models for Data Mining and Affective Computing (RAFDA), serves as an inclusive platform exploring the intricate intersections of foundation models, including LLMs, data mining, and affective computing. RAFDA represents a converging space uniting researchers focused on the applications, advancements, and implications of foundational models within the realms of data mining and affective computing.
At its core, RAFDA encapsulates interdisciplinary dialogue, spotlighting the innovative utilisation of cutting-edge foundational models for robust data mining practices. Simultaneously, it delves into the nuanced understanding and interpretation of affective computing, nurturing an inclusive forum for collaboration and knowledge exchange among researchers, practitioners, and industry experts. RAFDA also aims to be an international workshop to facilitate dynamic discussions among researchers in foundation models, NLP and data mining, fostering a collaborative environment for sharing groundbreaking research, novel methodologies, and innovative applications. RAFDA seeks to explore the diverse applications of foundation models, including LLMs, across academic and industrial domains, paving the way for envisioning future advancements and directions within the realm of data mining and AI research.

Important Dates

Special notes: We offer two submission rounds for the RAFDA workshop, providing additional opportunities for authors. For authors from China, we strongly recommend meeting the first-round submission deadline (Jan 17 2024) due to the potential three-month period required for visa processing to enter Taipei. We are pleased to confirm that the proceedings of our RAFDA workshop will be published by Springer in the Lecture Notes in Artificial Intelligence (LNAI) series. To attract more high-quality papers, we have extended the deadline (Feb 7 2024)for the second round of submissions by one week. The new deadline is Feb 14 2024.

Importance: To ensure sufficient time for quality checking and timely publication of our workshop proceedings by Springer, we kindly request that you submit your Camera Ready Paper by March 8, 2024. This will allow us to review the paper thoroughly and make any necessary revisions to meet the publishing deadlines and ensure availability before the conference event. Thank you for your cooperation and timely submission.

* 2nd round submissions(All paper deadlines are 23:59 Pacific Standard Time (PST))

Submission Deadline

February 14
2024

Acceptance Notification

February 28
2024

Camera Ready Paper

March 8
2024

Scope & Topics

  • Techniques for utilizing foundation models (like LLMs) for data preprocessing and feature extraction in mining large datasets.
  • Foundation models for anomaly detection, clustering, and classification in data mining.
  • Mining unstructured data using foundation models for improved pattern recognition and data analysis.
  • Sentiment analysis using foundation models in diverse applications and industries.
  • Emotion recognition and affective computing through foundation models.
  • Understanding human emotions through large-scale textual data and deep learning techniques.
  • Investigating affective computing methods integrating linguistic and computational approaches, aiming to understand and interpret emotions, sentiments, and affective states through machine learning and foundation models, e.g., LLMs, or other data mining techniques.
  • Interpretability and explainability of foundation models in affective computing.
  • Utilizing foundation models, e.g., LLMs for large-scale data analysis and pattern recognition.
  • Ethical considerations in deploying foundation models for data mining and affective computing.
  • Addressing biases and fairness issues in foundation model-based applications.
  • Societal impact and privacy concerns related to the utilization of LLMs.
  • Adaptation and fine-tuning of foundation models, e.g., LLMs for domain-specific affective data.
  • Applications of foundation models, e.g., LLMs in sentiment-aware decision-making processes.
  • Semantic understanding and context-based affective analysis.
  • Techniques to interpret and explain the predictions of foundation models in data mining and affective computing.
  • Explaining the decision-making process of foundation models for better transparency and trust.
  • Encompassing language modeling, text generation, sentiment analysis, summarization, and various aspects of data mining.
  • Architecture and functioning of foundation models, such as LLMs (e.g., ChatGPT).
  • Integration of visual, textual, and phonetic information using foundation models for improved data understanding and analysis.
  • Multimodal sentiment analysis and affective computing with fusion of diverse data sources.
  • Conversational AI and text generation tasks.
  • Applications of foundation models, e.g., LLMs in healthcare, finance, customer service, education, and other industry-specific domains.
  • Case studies and success stories showcasing the impact of foundation models in various fields.
  • Hands-on sessions for participants to engage with foundation models, e.g., LLMs, experiment with applications, and brainstorm future directions in the field.

Paper Submission

Submitted papers must be in English. The Program Committee will undertake a double-blind review of all papers based on their technical merit, relevance to the field of data mining, originality, significance, and clarity. Every submitted paper should include an abstract of no more than 200 words and should not exceed 12 single-spaced pages using a 10pt font size, including references, appendices, etc. Authors are instructed to follow the Springer LNCS/LNAI manuscript submission guidelines for their submissions.
All papers must be electronically submitted via the CMT paper submission system (https://cmt3.research.microsoft.com/RAFDA2024), and only in PDF format. Although authors can submit supplementary material in a separate PDF file, the reviewers are not obliged to consider it. Any papers failing to adhere to the Submission Policy will be rejected without review. Submitting a paper signifies that if the paper is accepted, at least one author will undertake regular registration and present the paper.
All submissions to RAFDA must be original work and not be under review or published in any other conference or journal. Submissions of papers must conform to the double-blind review policy, requiring the removal of any information identifying the author(s) from the main manuscript and any supplementary files. In discussing previous work, authors should reference their own studies in the third person and include all appropriate citations.

Formatting Template

All manuscripts should strictly adhere to the aforementioned format for preparation and submission. Deviation from this format may result in the disqualification of the paper from the RAFDA workshop of the conference.

Journal Publication

Extended versions of selected workshop papers will be published in either Cognitive Computation or IEEE ACSA.

Organizing Committee

Prof. Zhaoxia WANG
School of Computing and Information Systems
Singapore Management University
zxwang@smu.edu.sg
Prof. Erik CAMBRIA
School of Computer Science and Engineering
Nanyang Technological University
cambria@ntu.edu.sg
Prof. Bing LIU
School of Computer Science
University of Illinois at Chicago
liub@uic.edu
Dr. Boon Kiat QUEK
Social and Cognitive Computing Department
Institute of High Performance Computing, A*STAR
quekbk@ihpc.a-star.edu.sg
Dr. Seng-Beng HO
Social and Cognitive Computing Department
Institute of High Performance Computing, A*STAR
hosb@ihpc.a-star.edu.sg

Contact Information

Workshop Co-Chairs of RAFDA, PAKDD2024
rafda@socialopinionanalytics.com
rafda.pakdd2023@gmail.com

Program

RAFDA 2024 (A workshop of PAKDD 2024, May 7th, Taipei)


GO TO TOP

Each presentation is allocated 20 mins (15-min presentation + 5-min Q&A session). However, please note that the program is just indicative: if a presenter is missing, we'll go ahead with the next talk. All times are in Western Standard Time (GMT+8).

Two laptops will be available: one at the podium for your presentation and the other at the control console, managed by conference staff. All our presenters, please kindly save your presentation slides on the laptop at the podium before 8:45 am in the morning and 1:20 pm in the afternoon. This will streamline the process and ensure smooth operations.

CONFERENCE VENUE (Location)

Taipei International Convention Center (No. 1, Section 5, Xinyi Road, Xinyi District, Taipei City)
Session Name: Workshop RAFDA
Room: TICC 2F R201D
The conference room is available for use starting at 08:00 AM.

PROGRAM

09:00-09:05 Welcoming and introduction (RAFDA 2024 Chairs )
09:05-09:35 Keynote Talk (Morning): Sentiment Analysis Beyond Public Opinions (Yi Chen)
09:35-09:55 Explainable AI for Stress and Depression Detection in the Cyberspace and Beyond (Erik Cambria)
09:55-10:15 InteraRec: Interactive Recommendations Using Multimodal Large Language Models (Theja Tulabandhula)

☕   COFFEE BREAK

10:45-11:05 Deep Learning-based Vocal Separation for Audio into Music Sheet Conversion (Nicole Teo, Ezekiel Ghe, and Kevan Oktavio)
11:05-11:25 From Tweets to Token Sales: Assessing ICO Success through Social Media Sentiments (Donghao Huang)
11:25-11:45 Construction of Academic Innovation Chain Based on Multi-level Clustering of Field Literature (Wei Cheng)
11:45-12:05 Enhancing Child Safety: Multimodal Approach for Video Sentiment Analysis in Online Environments (Yee Sen Tan)

🍝  LUNCH BREAK

13:30-14:00 Keynote Talk (Afternoon): Bridging Trust: Combating Fakes and Frauds with Robust Graph Learning (Cheng-Te Li )

14:00-14:20 Evaluation of Orca 2 against other LLMs for Retrieval Augmented Generation (Donghao Huang)
14:20-14:40 Research on Dynamic Community Detection Method Based on Multi-dimensional Feature Information of Community Network (Kui Hu)
14:40-15:00 Toward Interpretable Graph Classification via Concept-Focused Structural Correspondence (Tien-Cuong Bui)
15:00-15:20 Enhanced Graph Neural Network for Session-Based Recommendation with Static and Dynamic Information (Kai Zheng)
15:20-15:30 Final Remarks and Group Photo Session: RAFDA 2024 (Everyone)
Group Photo of RAFDA 2024

KEYNOTE SPEAKER (Morning)
Prof Yi Chen
Prof Yi Chen

Prof Yi Chen holds the Martin Tuchman '62 Chair and is a Professor of Business Data Science in the Martin Tuchman School of Management, with a joint appointment in the Computer Science Department, at the New Jersey Institute of Technology. Her current research focuses on machine learning applications for users, encompassing user engagement, recommender systems, privacy, health, and well-being. She has served as the Associate Editor for TKDE, PVLDB, IJOC, and was the General Chair for SIGMOD 2012. She is a recipient of the PVLDB Distinguished Associate Editor Award, the Peter Chen Big Data Young Researcher Award, Google Research Awards, IBM Faculty Awards, and an NSF CAREER Award.

ABSTRACT

Sentiment analysis uses natural language processing and machine learning techniques to assess the emotional tone in texts and classify them as positive, negative, or neutral. It is widely used to analyze user-generated content, thus offering valuable insights into public opinions on products, services, or events. However, sentiment analysis can extend beyond the analysis of public opinions, offering the potential for analyzing a wider range of content across a broad spectrum of applications. One novel application of sentiment analysis lies in the medical domain. As a case study, we propose drawing an analogy between cancer biomarker extraction from pathology reports and aspect-based sentiment analysis in product reviews. This allows for leveraging and adapting a wealth of sentiment analysis literature to the medical domain, bringing significance to cancer diagnosis, prognosis, and treatment planning.

KEYNOTE SPEAKER (Afternoon)
Prof Cheng-Te Li
Prof Cheng-Te Li

Prof Cheng-Te Li is currently Full Professor at the Department of Computer Science and Information Engineering, National Cheng Kung University (NCKU) in Tainan, Taiwan. He earned his Ph.D. degree in 2013 from the Graduate Institute of Networking and Multimedia at National Taiwan University. Prior to joining NCKU, Dr. Li served as an Assistant Research Fellow at CITI, Academia Sinica, from 2014 to 2016. Focusing on Machine Learning and Data Mining, Dr. Li's research explores their applications in Social Networks, Social Media, Recommender Systems, and Natural Language Processing. His work has been featured at premier conferences such as KDD, TheWebConf (WWW), ICDM, CIKM, SIGIR, IJCAI, ACL, EMNLP, and NAACL. Recently, his group has presented lecture-style tutorials on Graph Neural Networks at top conferences, including WWW, IEEE ICDE, and ACML. Dr. Li leads the Networked Artificial Intelligence Laboratory (NetAI Lab) at NCKU.

ABSTRACT

In the digital era, where misinformation and fraud proliferate, robust graph learning emerges as a critical tool for restoring trust. This talk delves into the advances graph learning for counteracting the complexities of fake news, rumors, and fraudulent activities across various domains. Leveraging the robust graph neural networks learned from transactions, text, and tabular datasets, this talk will show how innovations in generative methods, such as data augmentations and synthesis, and self-supervised learning, such as contrastive and curriculum learning, are reshaping the landscape of trustworthiness in digital and physical marketplaces. These advanced graph models not only enhance the detection and mitigation of deceitful practices but also pave the way for more transparent, reliable information exchange. The presentation will highlight graph learning's effectiveness in building robust ecosystems immune to falsehoods, focusing on applications in rumor and customs fraud detection, AI-driven food safety, and social media privacy. In doing so, it will present a strategy for employing powerful graph models to navigate and neutralize the challenges brought on by disinformation and deceit, showcasing a committed pathway toward upholding truth in our interconnected society.


RAFDA 2024