Yimin Shi

Yimin Shi

Auto-generated profile, based on publicly available information

Yimin Shi

National University of Singapore, Singapore, Singapore

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 July 2025
Verified
Conference Paper
Open access
YSYimin Shi
Yimin Shi
YFYang Fei
Yang Fei
SZShiqi Zhang
Shiqi Zhang
HWHaixun Wang
Haixun Wang
XXXiaokui Xiao
Xiaokui Xiao


18 May 2026 in Proceedings of the ACM on Management of Data
Verified
Journal Article
Open access
XCXingguang Chen
Xingguang Chen
YSYimin Shi
Yimin Shi
XXXiaokui Xiao
Xiaokui Xiao
Entity matching is a fundamental task in data management, supporting applications such as product matching in e-commerce, linking scholarly records, unifying location data, and cross-source fact-checking. Its generalized form, known as generalized entity matching (GEM), extends the challenge to heterogeneous sources across structured, semi-structured, and unstructured data, where schema mismatch, structural variation, and domain drift often make supervision costly or infeasible. Existing approaches either rely heavily on labeled data or fail to generalize across domains, lacking the ability to use both structural and content information at scale. We present GLEAM, an end-to-end generalized entity matching framework that dynamically adapts to data structure and domain characteristics without requiring labeled pairs. GLEAM integrates (i) a structure- and content-aware blocking module that constructs a weighted heterogeneous graph encoding schema context and semantic similarity, guided by a lightweight LLM-based attribute importance estimator, (ii) an adaptive connector that models candidate score distributions via a two-component GMM and performs Bayesian updates using online feedback from matching, enabling entity-specific early stopping, and (iii) a hierarchical reasoning matching module that infers domain and attribute hierarchies from a few examples and performs comparative selection over candidate sets with domain-aware prompts. Across a diverse set of heterogeneous datasets, GLEAM consistently achieves strong end-to-end F1 while reducing unnecessary LLM calls. It outperforms leading unsupervised and LLM-based baselines and remains competitive with supervised methods, while achieving up to a 25.7% improvement in F1. We further provide probabilistic analyses of the connector's thresholding behavior and update rules, along with ablation studies on graph weighting, adaptivity, and hierarchical prompting.

26 December 2017
Verified
Journal Article
Closed access
JTJ Tang
J Tang
SZS Y Zhang
S Y Zhang
FLF J Lu
F J Lu
WQW H Qin
W H Qin
YSY S Shi
Y S Shi
Objective: To observe the effect of low dose naloxone combinewith ropivacaine for supraclavicular brachial plexus block. Methods: Seventy patients undergoing elective upper limb surgery were randomly divided into two groups, ropivacaine group (Group R, n=35) and naloxone group (Group N, n=35). An ultrasound guided technique was used in both two groups.The onset and duration time of sensory and motor blockade, visual analog score(VAS)of 3, 6, 12, 18, 24 h postoperatively, time of first request fordezocine, total amount of dezocine needed, incidence of nausea and vomiting postoperatively(PONV) and patients' satisfaction score for analgesia in 24 h after surgery were measured.At the same time, blood samples were taken before anesthesia, 6 h, 24 h after operation for inspecting the concentration of β-endorphin(β-EP)in plasma. Results: The duration of sensory and motor blockade, time of first request for dezocine in Group N were 736.0(713.5, 836.5), 514.5(491.3, 572.8), 708.5(683.2, 877.0)min, which were all prolonged compared to Group R(522.0(469.5, 606.5), 401.0(370.0, 458.5), 570.0(435.0, 618.5)min)(Z=-6.844, -6.758, -6.700, all P<0.01). The 6, 12, 18 h postoperatively VAS of Group N were 0, 5.0(3.0, 5.8), 5.0(5.0, 6.0)point. Among which the 6, 12 h postoperatively VAS of Group N were lower than that of Group R(1.0(1.0, 3.5), 6.0(6.0, 7.0)point)(Z=-6.596, -4.864, all P<0.01), while the 18 h postoperatively VAS was higher than that of Group R (5.0(4.0, 5.0)point)(Z=-2.603, P<0.01). Total amount of dezocine needed in Group N in 24 h after surgery was 7.5(5.0, 10.0)mg, which was lower than that of Group R(10.0(10.0, 15.0)mg)(Z=-3.449, P<0.01). The incidence of PONV after surgery in Group N was 21.9%, which was lower than that of Group R(45.5%)(χ(2)=4.034, P<0.05). Ptients' satisfaction score for analgesia in 24 h after surgery in Group N was 8.0(7.0, 8.0)point, which was higher than that of Group R(7.0(6.0, 7.0)point)(Z=-3.509, P<0.01). At 6 h postoperatively , the concentration of plasma β-EP in Group N was(113.34±12.36)μg/L, lower than that of Group R((147.14±11.65)μg/L)(t=-7.694, P<0.01). Conclusion: Low dose naloxone combine with ropivacaine for supraclavicular brachial plexus block, prolong the duration of sensory and motor blockade without affecting the onset time.

8 August 2026
Verified
Conference Paper
Open access
SHSilan Hu
Silan Hu
SZShiqi Zhang
Shiqi Zhang
YSYimin Shi
Yimin Shi
XXXiaokui Xiao
Xiaokui Xiao
Generative Engine Marketing (GEM) is an emerging ecosystem for monetizing generative engines, such as LLM-based chatbots, by seamlessly integrating relevant advertisements into their responses. At the core of GEM lies the generation and evaluation of ad-injected responses. However, existing benchmarks are not specifically designed for this purpose, which limits future research. To address this gap, we propose GEM-Bench, the first comprehensive benchmark for ad-injected response generation in GEM. GEM-Bench includes three curated datasets covering both chatbot and search scenarios, a metric ontology that captures multiple dimensions of user satisfaction and engagement, and several baseline solutions implemented within an extensible multi-agent framework. Our preliminary results indicate that, while simple prompt-based methods achieve reasonable engagement such as click-through rate, they often reduce user satisfaction. In contrast, approaches that insert ads based on pre-generated ad-free responses help mitigate this issue but introduce additional overhead. These findings highlight the need for future research on designing more effective and efficient solutions for generating ad-injected responses in GEM.

8 August 2026
Verified
Conference Paper
Open access
YSYimin Shi
Yimin Shi
CWClarice Wang
Clarice Wang
HWHaixun Wang
Haixun Wang
XXXiaokui Xiao
Xiaokui Xiao
Many real-world conversational settings for knowledge discovery, including podcasts, hiring screens, and marketplaces, need a purpose-driven understanding of a person. For this, the most effective path is to ask the Next Best Question (NBQ) at each turn: the question with the highest expected information gain given what has already been learned and the conversation's goal. We propose the NBQ framework, a plug-and-play smart asker that, given any topic, seeds a diverse pool of hundreds to thousands of candidate questions, maintains a compact, continuously updated user state that tracks coverage and confidence, selects the next question after every answer to maximize incremental value within the prespecified turn budget, and finally distills the unstructured Q&A dialogue into a structured, vector-based user profile ready for downstream mining tasks. As a demanding application, we instantiate NBQ for reciprocal matchmaking, where ''reciprocal'' means compatibility must be mutual: i.e., a match between two people is valid only when one fits the other's preferences, and vice versa. Therefore, each person is modeled with two representations: (i) who they are (self-description) and (ii) whom they prefer (counterpart preferences). NBQ asks questions to refine both vectors and thereby increase the probability of a successful match. To scale matching to real-world platforms with millions of concurrent users and continuously updated profiles, we introduce QuickMatch, an efficient retrieval layer that recasts reciprocal matching from quadratic pairwise scoring to approximate vector search. With modest storage overhead, QuickMatch updates each user's top matches in real time. Compared with random or conventional generative questioning baselines, NBQ improves the quality of user profiling by up to 13.6% and 14.0% in terms of AC@T and AR@T. Meanwhile, QuickMatch accelerates retrieval by up to 22.9× while maintaining a high recall of up to 0.989.