1st Workshop on Multimodal Search and Recommendations (CIKM MMSR ‘24) Date: October 25, 2024 (Full day workshop) Venue: ACM CIKM 2024 (Boise, Idaho, United States) Website: https://cikm-mmsr.github.io/ Overview: The advent of multimodal LLMs like GPT-4o and Gemini has significantly boosted the potential for multimodal search and recommendations. Traditional search engines rely mainly on textual queries, supplemented by session and geographical data. In contrast, multimodal systems create a shared embedding space for text, images, audio, and more, enabling next-gen customer experiences. These advancements lead to more accurate and personalized recommendations, enhancing user satisfaction and engagement. Topics of interest include, but are not limited to: Cross-modal retrieval techniques Strategies for efficiently indexing and retrieving multimodal data. Approaches to ensure cross-modal retrieval systems can handle large-scale data. Development of metrics to measure similarity across different data modalities. Applications of Multimodal Search and Recommendations to Verticals (e.g. E-commerce, real estate) Implementing and optimizing image-based product searches. Creating multimodal conversational systems to enhance user experience and make search more accessible. Utilizing AR to enhance product discovery and user interaction. Leveraging multimodal search for efficient customer service and support. User-centric design principles for multimodal search interfaces Best practices for designing user-friendly interfaces that support multimodal search. Methods for evaluating the usability of multimodal search interfaces. Personalizing multimodal search interfaces to individual user preferences. Ensuring multimodal search interfaces are accessible to users with disabilities. Ethical Considerations and Privacy Implications of Multimodal Search and Recommendations Strategies for ensuring user data privacy in multimodal applications. Identifying and mitigating biases in multimodal algorithms. Ensuring transparency in how multimodal results are generated and presented. Approaches for obtaining and managing user consent for using their data. Modeling for Multimodal Search and Discovery Multi-modal representation learning Utilizing GPT-4o, Gemini, and other advanced pre-trained multimodal LLMs Dimensionality reduction techniques to reduce complexity of multimodal data. Techniques for fine-tuning pre-trained vision-language models. Developing and standardizing metrics to evaluate the performance of vision-language models in multimodal search. Submission Instructions: All papers will be peer-reviewed by the program committee and judged based on their relevance to the workshop and their potential to generate discussion. Submissions must be in PDF format, following the latest CEUR single column format. For instructions and LaTeX/Overleaf/docx templates, refer to CEUR’s submission guidelines ( https://ceur-ws.org/HOWTOSUBMIT.html#CEURART), reading up to and including the “License footnote in paper PDFs” section. Use Emphasizing Capitalized Style for Paper Titles. Submissions must describe original work not previously published, not accepted for publication, and not under review elsewhere. All submissions must be in English. The workshop follows a single-blind review process and does not accept anonymous submissions. At least one author of each accepted paper must register for the workshop and present the paper. Long paper limit: 15 pages. Short paper limit: 8 pages. References are not counted in the page limit. Submit to CIKM MMSR’24: https://openreview.net/group?id=ACM.org/CIKM/2024/Workshop/MMSR The deadline for paper submission is August 10, 2024 (23:59 P.M. GMT) Contact: Aditya Chichani (Walmart, University of California Berkeley) E-mail: aditya_chichani@berkeley.edu