Time: 12:30 PM, June 4, 2026(Thursday)
Location: Room 1100, 11th Floor, Lyu Dalong Building, Tsinghua University
1. Xiaoang Wan’s Research Group
Presentation Topic:The Belief Conversion Asymmetry: A Cue-to-Belief Framework for Understanding Biased Reliance on AI Advisors
Presenter: Mengyin Liu
Abstract：
AI advisory systems are increasingly deployed in high-stakes decision-making, yet how individuals convert AI competence information into calibrated reliance behavior remains poorly understood. Across three studies using a judge-advisor paradigm (N = 1,080), we examined how performance metrics drive advice-taking from AI versus human advisors at behavioral and metacognitive levels.
Study 1 revealed a framing-dependent asymmetry in competence label-sensitive calibration: accuracy labels calibrated both advice adoption and confidence change for human but not AI advisors, whereas error-rate labels eliminated competence label-sensitive calibration across both advisor types.
Study 2 demonstrated a framing-independent pattern in which internally held competence beliefs symmetrically calibrated both outcomes across AI and human advisors.
Study 3 found that participants tracked AI and human advisors’ performance with comparable sensitivity but formed less stable accuracy estimates for AI advisors. After controlling for estimate stability, newly acquired competence estimates produced symmetric effects across advisor types, with both predicting confidence change, though insufficient to predict advice adoption. This finding suggests that once competence beliefs are stabilized, AI input can be weighted equivalently to human input.
Collectively, our findings locate the AI-human asymmetry in competence-based calibration at the stage where external performance cues are converted into internal competence beliefs, not where those beliefs guide advice weighting. Effective human-AI collaboration thus requires a shift from merely making AI competence visible to helping users internalize AI performance information as stable and actionable competence beliefs.
2. Meihong Zheng’s Research Group
Presentation Topic: Item Matching or Structural Representation? The Mechanism of Duration Organization in Working Memory
Presenter: Yunxin Zhao
Abstract：
How multiple durations are organized in working memory and utilized for subsequent judgments remains an unresolved, critical question in the field of time cognition. It also relates to a more fundamental question: does working memory merely preserve discrete items, or is it capable of extracting structures from continuous experiences to form internal organizations? Traditional views typically conceptualize duration memory as a collection of independent item traces; that is, during recognition, individuals match a probe duration against encoded durations one by one to judge whether it has appeared before. However, this perspective overlooks a crucial possibility: when multiple durations are sequentially encoded within the same context, what working memory preserves may not be isolated durations, but rather temporal information organized around the global structure of the sequence.
Through two duration recognition experiments combined with computational modeling, this study investigates the multi-level temporal information upon which duration recognition relies. The results indicate that recognition judgments are shaped not only by local item similarity, but also jointly by the global range at the task level and the central tendency information at the trial level. Notably, unpresented probe durations located at the center of the learned sequence could induce stable false recognition. This suggests that duration recognition does not depend entirely on whether a specific duration actually occurred, but is also systematically influenced by the overall structure of the sequence. Further analysis revealed that basic temporal precision primarily explains an individual's utilization of local item information, yet it cannot reliably account for their sensitivity to global range and sequence-center information. After controlling for temporal precision, participants' reliance on sequence-center information and local item similarity still exhibited a negative correlation, suggesting that duration recognition may involve two distinct processing orientations: a holistic processing approach oriented toward summary structures, and a local matching approach oriented toward specific items.
Based on these findings, this study proposes a structural duration organization perspective. This view posits that multiple durations are not merely a collection of independent traces in working memory, but may instead be organized into a multi-level structural representation that includes global boundaries, sequence centers, and local item information. This study challenges the traditional understanding that simplifies duration recognition into item matching. It reveals that temporal working memory does not simply passively store and match discrete durations during recognition judgments; rather, it can extract structural information from continuous temporal experiences and support subsequent judgments through multi-level representational retrieval. This research not only expands our understanding of the mechanisms underlying duration recognition, but also provides new evidence revealing how working memory transforms continuous experience into structured internal models.
