Answers are grounded in this paper and cite their sources.
Large language models (LLMs) and other foundation models have achieved great success, and many KDD tasks have been solved with them. What research directions worth attention for future? This article presents three directions in my mind. First, how to make predictions made by learning models trustable, reliable and accountable? Second, how to build AI model management systems enabling model identification, reuse and even assembly, without touching raw data? Third, when a learning model predicts that something undesired is going to happen, how to avoid undesired future?
No takes yet. Share an insight, caveat, or question.
Zhi‐Hua Zhou (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: