KDD OARS Workshop

Introduction

The international workshop on Online and Adaptive Recommender Systems (OARS) will serve as a platform for publication and discussion of OARS. This workshop will bring together practitioners and researchers from academia and industry to discuss the challenges and approaches to implement OARS algorithms and systems, and improve user experiences by better modeling and responding to user intent.

Many recommender systems deployed in the real world rely on categorical user-profiles and/or pre-calculated recommendation actions that stay static during a user session. Recent trends suggest that recommender systems should model user intent in real time and constantly adapt to meet user needs at the moment or change user behavior in situ. In addition, various techniques have been proposed to help recommender systems adapt to new users, items, or behaviors. Some strategies to build “adaptive” recommenders include:

Systems for online training, e.g. updating the parameters of a pre-trained model to a new user.
Feature-based systems that handle cold-start scenarios, and can gracefully adapt to a combination of cold- and warm users/items.
Systems that avoid modeling users at all (e.g. session-based recommenders that directly learn from item interactions without needing user terms)
Systems that adapt to new behaviors through RL or other adaptive learning algorithms.

We invite submission of papers and posters of two to ten pages (including references), representing original research, preliminary research results, proposals for new work, position, and opinion papers. All submitted papers and posters will be single-blind and will be peer reviewed by an international program committee of researchers of high repute. Accepted submissions will be presented at the workshop.

Topics of interest include, but are not limited to:

Foundation and LLM models in RecSys
- Conversational and interactive RecSys using LLMs
- Personalized agents for RecSys with LLMs
- Enhance product understanding and representation using foundation and LLM models
- Improve the generalization and relevancy of RecSys via LLM and Multi-modal foundation models
- Elevate recommendation experiences with foundation models
- Improve the explanation and reasoning of RecSys via LLMs
- Evaluation of RecSys using LLMs
- Scaling foundation and LLM models in RecSys
- Embedding evaluation for embedding, ranking, and other services
Novel algorithms and paradigms
- online and adaptive neural recommender
- reinforcement learning (on-policy, off-policy, offline RL, and other relevant subfields)
- online/streaming learning
- interactive and conversational recommender
- extreme classification
- graph recommender
Applications
- product recommendations
- content recommendations (news, music, movie, video, etc.)
- ads recommendations
- fashion and decor recommendation
- job recommendation
- intervention/behavior change/healthy life-style recommendations
User modeling and representations
- implicit and explicit user intent modeling
- dynamic user intent modeling
- visual/style/taste modeling
- combination of in-session intent with long term user interest
- incorporation of knowledge graph
- representation learning
Architecture and infrastructure
- scalability of neural methods for large scale real-time recommendations
- steaming and event-driven processing infrastructures
Evaluation and explanation methodologies
- evaluation, comparison, explanation of OARS for a recommendation task
- off-policy and counterfactual evaluation
Social and user impact
- UX for OARS
- welfare and objectives of OARS (CTR, dwell-time, diversity, multi-objectives, long term objectives)
- privacy and ethics considerations

Call for Papers

Download CFP

All papers will be peer reviewed (single-blind) by the program committee and judged by their relevance to the workshop, especially to the main themes identified above, and their potential to generate discussion.

All submissions must be formatted according to the latest ACM SIG proceedings template (two column format). One recommended setting for Latex file of manuscript is: \documentclass[sigconf, anonymous, review]{acmart}. Submissions must describe work that is not previously published, not accepted for publication elsewhere, and not currently under review elsewhere. All submissions must be in English.

Please note that at least one of the authors of each accepted paper must register for the workshop and attend in person to present the paper during the workshop.

Submissions to CIKM OARS workshop should be made at easychair page.

Key Dates

Submissions Due	~~July 29~~Aug 16
Notification	Aug 30
Camera Ready Version Due	Sept 30
Workshop Day	Oct 25

Feedback and stay updated to the workshop

Schedule

Time	Talk
9:00-9:15 PT	Openning
9:15-10:00 PT	Invited Talk 1 Powering Creative Asset Recommendations with Hybrid Multi-Modal Search Tracy Holloway King, Adobe Creating marketing and social media content is a challenge for both expert and novice creators. This content is usually multi-modal, involving multiple images, text, video, and audio. Providing the creator with contextual recommendations based on the current project state and its latent intent can spark new ideas and help creators discover relevant content and complete their projects more quickly. In this talk, I discuss how Adobe is leveraging symbolic and embedding-based intent understanding and a domain-specific knowledge graph to understand the creator's project, determine what type of assets are most relevant (e.g. images, icons, backgrounds), and leverage multi-modal search to provide a series of asset recommendations that complement the project both in intent and aesthetics.
10:00-10:15 PT	Contributed Talk 1 Real-time Event Joining in Practice With Kafka and Flink Srijan Saket, Vivek Chandela and Danish Kalim
10:15-10:30 PT	Contributed Talk 2 A Good State Estimator Can Yield A Simple Recommender: A Reinforcement Learning Perspective Dilina Rajapakse and Douglas Leith
10:30-11:00 PT	Coffee Break
11:00-11:45 PT	Invited Talk 2 From Collaborative Filtering to LLMs: Building a Cutting-Edge Recommendation & Search Platform Tullie Murrell, Shaped AI This talk details Shaped's journey in developing a next-generation recommendation and search platform, tracing its evolution from traditional collaborative filtering to the adoption of powerful Large Language Models (LLMs). I'll delve into the challenges of building a multi-tenanted platform capable of serving diverse use cases at scale. Finally, I'll present future directions for generalized multi-domain RecSys, illustrating the transformative potential of large scale embedding models in shaping the next generation of recommendation and search systems.
11:45-12:30 PT	Invited Talk 3 TBD Tyler Derr, Vanderbilt University TBD
12:30-01:45 PT	Lunch Break
01:45-02:30 PT	Invited Talk 4 LLMs and RecSys: From Graph Structure Knowledge to Dynamic LLMs Routing Yanjie Fu, Arizona State University Large language models (LLMs) are effective for generating personalized and explainable recommendations from user and system behavioral data. While they excel in understanding contexts and languages, LLMs struggle with graph-structured data, which is often how user-item interactions are represented. This limits their ability to deliver accurate recommendations in complex scenarios. Additionally, given the computational cost and latency of various LLMs, recommending the most suitable LLM for queries presents a challenge. In this talk, we will firstly talk about graph structured LLM for recommendations, and then introduce MixLLM, a dynamic routing system that recommends the most suitable LLM for queries to balance among response quality, costs, and latency. Finally, we will conclude our talk and discuss future work.
02:30-02:45 PT	Closing remarks.