WSDM 2027 · Hong Kong
Call for WSDM Day Presentations
State of Data Mining & Web Search in the Agentic AI Era
Extended abstract2 pages
In ACM sigconf format, written in English.
Review processSingle-blind
Include all authors' names and affiliations.
PresentationOral & poster
The format will be decided after acceptance.
Overview
Artificial intelligence is rapidly evolving from systems that generate content or answer individual queries
into agents that search, reason, plan, interact with tools and environments, and carry out complex tasks on
behalf of users. Web search and data mining are fundamental to this transition. Agents must discover and
retrieve information, reason over heterogeneous and evolving data, learn from interactions, maintain and
update knowledge and memory, and make decisions grounded in reliable evidence.
At the same time, agentic systems are changing how people interact with the Web. Traditional search and
recommendation increasingly coexist with systems that synthesize information across sources, perform
multi-step exploration, invoke tools and services, and act on users' behalf. These developments raise
fundamental research questions concerning information access, personalization, evaluation, reliability,
safety, and the future of the Web ecosystem.
This year, WSDM Day aims to promote WSDM-aligned research on these emerging challenges and discuss the
“State of Data Mining & Web Search in the Agentic AI Era.”
Topics of Interest
The organizing committee invites submissions for oral and poster presentations on topics aligned with the
WSDM community, with a particular focus on the intersection of web search, data mining, recommendation, and
agentic AI. We welcome submissions related to, but not limited to, the following topics.
Search, Retrieval, and Knowledge for Agents
Agentic systems need to acquire information dynamically rather than rely solely on knowledge encoded
in their underlying models. Search, retrieval, ranking, and knowledge integration therefore become core
capabilities of agents operating in open and continuously changing environments.
Topics in this area include:
- Web search, information retrieval, and retrieval-augmented generation for agents
- Agentic search and multi-step information seeking
- Query generation, reformulation, and adaptive retrieval
- Ranking, filtering, and evidence selection for agentic systems
- Knowledge graphs and structured knowledge integration
- Freshness, provenance, attribution, and citation
- Multimodal and cross-lingual information access for agents
- Agents for exploratory and complex information-seeking tasks
Planning, Reasoning, Memory, and Learning
Agents must reason across multiple steps, maintain state, learn from interactions, and adapt their
strategies as tasks and environments evolve. Web-scale data and user interactions provide rich signals
for building and improving these capabilities.
Topics in this area include:
- Planning and reasoning for web-based agents
- Long-horizon and hierarchical task execution
- Short-term, long-term, and episodic memory
- Continual learning and self-improving agents
- Learning from user feedback and interaction data
- Tool use and API orchestration
- Multi-agent collaboration and coordination
- Personalization and user modeling for agents
Agents for Search, Recommendation, and the Web
Agentic AI is reshaping established WSDM problems including search, recommendation, advertising, and
online interaction. Conversely, decades of research in these areas provide important foundations for
building effective agentic systems.
Topics in this area include:
- Agentic recommender systems
- Conversational and interactive search agents
- Web and GUI agents
- Personalized agents and user assistants
- Agents for e-commerce, social platforms, and online services
- Human-agent collaborative information seeking
- Agent-mediated access to online information and services
- New interaction paradigms beyond traditional search and recommendation interfaces
Evaluation and Benchmarking of Agentic Systems
Evaluating agents presents challenges that go beyond conventional model evaluation. Successful
execution may require multiple steps, external information, changing environments, and interactions
whose quality cannot be captured by a single static metric.
Topics in this area include:
- Benchmarks and datasets for web and information-seeking agents
- Evaluation of long-horizon and multi-step tasks
- Evaluation of search, retrieval, reasoning, and action as an integrated system
- Online and interactive evaluation
- User-centered evaluation of agentic systems
- Robustness and generalization across tasks and environments
- Reproducibility and evaluation in dynamic web environments
- Efficiency, latency, and cost-aware evaluation
Trustworthy and Responsible Agentic AI
As agents gain the ability not only to retrieve information but also to make decisions and take
actions, questions of reliability, accountability, privacy, and user control become increasingly
important.
Topics in this area include:
- Reliability, factuality, and hallucination in agentic systems
- Grounding and verification of agent-generated conclusions and actions
- Privacy and security for web agents
- Adversarial attacks and manipulation of agents
- Bias, fairness, and transparency
- Explainability and interpretability of agent decisions
- Human oversight, intervention, and control
- Safe tool use and action execution
- Trust, provenance, and accountability in autonomous systems
The Agentic Web and Emerging Research Directions
The growth of autonomous agents may fundamentally alter the structure and economics of the Web itself.
Increasingly, online information and services may be consumed, filtered, and acted upon by agents rather
than directly by human users.
We particularly welcome forward-looking work addressing questions such as:
- How should search and recommendation evolve when the primary consumer may be an agent?
- How should web content and services be designed for both humans and agents?
- How will agents affect information discovery, content creation, and online ecosystems?
- What new protocols, representations, and infrastructure are needed for an agentic Web?
- How should agents interact with websites, platforms, APIs, and other agents?
- How might agentic systems reshape incentives, attribution, content quality, and information diversity on the Web?
- What new research problems will emerge as agents become persistent participants in online ecosystems?
Submission Guidelines
Submissions are invited in the form of 2-page extended abstract. The format of presentation, oral or poster,
will be decided after acceptance.
All submissions must adhere to the formatting requirements specified in the WSDM 2027 author guidelines.
Submissions must be self-contained and written in English.
Submissions will be reviewed in a single-blind manner by at least two reviewers and must include all
authors' names and affiliations. Submissions that fail to adhere to the submission guidelines or fall outside
the scope of relevant topics may be rejected without review.
Accepted contributions will be presented as oral or poster presentations during WSDM Day. Accepted papers
will be featured in the WSDM Day program and included in the conference proceedings, subject to the
publication requirements of WSDM 2027.
Dual-submission Policy
We welcome ongoing and unpublished work. We also welcome papers that are under review at the time of
submission.
Presentation
The presentation format will include short oral presentations and a poster session. At least one author of
each accepted presentation must register for the conference and present the work in person during WSDM Day
2027.
Important Dates
- Submission Deadline
- Nov 30, 2026
- Notification of Acceptance
- Dec 21, 2026
- Camera Ready
- Dec 31, 2026
- WSDM Day
- Feb 19, 2027
All deadlines are 11:59 pm Anywhere on Earth (AoE).
Submission
Submissions should be made through the WSDM 2027 submission system:
Microsoft CMT.
We look forward to your submissions and participation in WSDM Day 2027!
Contact
For questions or additional information, please contact the organizing committee at
wsdm2027.acm@polyu.edu.hk.
WSDM Day Chairs
