Haobo Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Haobo Wang has a total of 207 co-authors. Here are the top co-authors by number of shared publications:
- Junbo Zhao
Zhejiang University, Hangzhou, China
25 shared publications Gang Chen
Zhejiang University, Hangzhou, China
18 shared publicationsTianlei Hu
Hangzhou High-Tech Zone (Binjiang) Institute of Blockchain and Data Security
11 shared publications- Gang Chen
Zhejiang University
10 shared publications - Lei Feng
State Key Laboratory of Switching and Networking Technology, Beijing University of Posts and Telecommunications, Beijing, China
7 shared publications - Runze Wu
NetEase Fuxi AI Lab, Hangzhou, China
7 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 13126-13135
Beyond Contrastive: Generative Reranking for Multilingual Text-Based Person Retrieval
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 13126-13135
Open AccessBeyond Contrastive: Generative Reranking for Multilingual Text-Based Person Retrieval
- 15 October 2026
- Published by Springer Nature
- p. 105-120
SmartInvest: Multi-view Enterprise Ranking and LLM-Assisted Investment Analysis
- 15 October 2026
- Published by Springer Nature
- p. 105-120
SmartInvest: Multi-view Enterprise Ranking and LLM-Assisted Investment Analysis
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 6282-6291
Hypergraph-Based Unaligned Multi-View Clustering via Cluster-Aware Feature Extraction
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 6282-6291
Open AccessHypergraph-Based Unaligned Multi-View Clustering via Cluster-Aware Feature Extraction
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 6188-6199
FLaG: Fine-Grained Latent Grouping for Hallucination Detection
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 6188-6199
Open AccessFLaG: Fine-Grained Latent Grouping for Hallucination Detection
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 9125-9136
DMind Benchmark: Toward a Holistic Assessment of LLM Capabilities across the Web3 Domain
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 9125-9136
Open AccessDMind Benchmark: Toward a Holistic Assessment of LLM Capabilities across the Web3 Domain
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 7401-7411
Reinforcement Learning with Verbalized Probabilities for LLM Classification
While Large Language Models (LLMs) excel at many reasoning tasks, their native inability to produce calibrated, multi-class probability distributions limits their use in high-stakes Web applications like content moderation and fraud detection. Existing methods to elicit probabilities from LLMs either sacrifice their crucial Chain-of-Thought (CoT) reasoning capabilities or suffer from poor calibration. To address this, we introduce a new paradigm, Verbalized Probability Distribution, and a novel training framework, RLVP (Reinforcement Learning with Verbalized Probabilities). RLVP fine-tunes an LLM to generate both an interpretable CoT and a complete, verbalized probability distribution. We overcome the ''insufficient reward granularity'' problem in standard Reinforcement Learning (RL) for classification by using soft probabilities from expert tabular models as a dense reward curriculum. Through large-scale joint training on 169 tabular tasks, we demonstrate that a single RLVP-trained model can surpass a strong, task-specific XGBoost baseline on up to 55% of tasks. More importantly, the trained model achieves state-of-the-art few-shot performance on unseen, heterogeneous Web benchmarks that mix structured data with free text, achieving performance comparable to or superior than expert models trained on the same limited data. This showcases a strong capability for generalization and knowledge transfer to complex Web data. Our work presents a viable path toward building general-purpose, probabilistically-sound, and interpretable foundation models for the Web.
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 7401-7411
Open AccessReinforcement Learning with Verbalized Probabilities for LLM Classification
While Large Language Models (LLMs) excel at many reasoning tasks, their native inability to produce calibrated, multi-class probability distributions limits their use in high-stakes Web applications like content moderation and fraud detection. Existing methods to elicit probabilities from LLMs either sacrifice their crucial Chain-of-Thought (CoT) reasoning capabilities or suffer from poor calibration. To address this, we introduce a new paradigm, Verbalized Probability Distribution, and a novel training framework, RLVP (Reinforcement Learning with Verbalized Probabilities). RLVP fine-tunes an LLM to generate both an interpretable CoT and a complete, verbalized probability distribution. We overcome the ''insufficient reward granularity'' problem in standard Reinforcement Learning (RL) for classification by using soft probabilities from expert tabular models as a dense reward curriculum. Through large-scale joint training on 169 tabular tasks, we demonstrate that a single RLVP-trained model can surpass a strong, task-specific XGBoost baseline on up to 55% of tasks. More importantly, the trained model achieves state-of-the-art few-shot performance on unseen, heterogeneous Web benchmarks that mix structured data with free text, achieving performance comparable to or superior than expert models trained on the same limited data. This showcases a strong capability for generalization and knowledge transfer to complex Web data. Our work presents a viable path toward building general-purpose, probabilistically-sound, and interpretable foundation models for the Web.
- 3 March 2026
- Published by Springer Nature in Machine Learning
- Vol. 115 (3) , 53
LADA: Label Disambiguation and Domain-Aware Learning for Domain Generalization
- 3 March 2026
- Published by Springer Nature in Machine Learning
- Vol. 115 (3) , 53
LADA: Label Disambiguation and Domain-Aware Learning for Domain Generalization
- Open Access
A Syllogistic Probe: Tracing the Evolution of Logic Reasoning in Large Language Models
Human logic has gradually shifted from intuition-driven inference to rigorous formal systems. Motivated by recent advances in large language models (LLMs), we explore whether LLMs exhibit a similar evolution in the underlying logical framework. Using existential import as a probe, we for evaluate syllogism under traditional and modern logic. Through extensive experiments of testing SOTA LLMs on a new syllogism dataset, we have some interesting findings: (i) Model size scaling promotes the shift toward modern logic; (ii) Thinking serves as an efficient accelerator beyond parameter scaling; (iii) the Base model plays a crucial role in determining how easily and stably this shift can emerge. Beyond these core factors, we conduct additional experiments for in-depth analysis of properties of current LLMs on syllogistic reasoning.
Open AccessA Syllogistic Probe: Tracing the Evolution of Logic Reasoning in Large Language Models
Human logic has gradually shifted from intuition-driven inference to rigorous formal systems. Motivated by recent advances in large language models (LLMs), we explore whether LLMs exhibit a similar evolution in the underlying logical framework. Using existential import as a probe, we for evaluate syllogism under traditional and modern logic. Through extensive experiments of testing SOTA LLMs on a new syllogism dataset, we have some interesting findings: (i) Model size scaling promotes the shift toward modern logic; (ii) Thinking serves as an efficient accelerator beyond parameter scaling; (iii) the Base model plays a crucial role in determining how easily and stably this shift can emerge. Beyond these core factors, we conduct additional experiments for in-depth analysis of properties of current LLMs on syllogistic reasoning.
- 20 January 2026
- Published by Springer Nature in World Wide Web
- Vol. 29 (2) , 16
Toward real-world Table Agents: capabilities, workflows, and design principles for LLM-based table intelligence
- 20 January 2026
- Published by Springer Nature in World Wide Web
- Vol. 29 (2) , 16
Toward real-world Table Agents: capabilities, workflows, and design principles for LLM-based table intelligence
- Open Access
An Invariant Latent Space Perspective on Language Model Inversion
Language model inversion (LMI), i.e., recovering hidden prompts from outputs, emerges as a concrete threat to user privacy and system security. We recast LMI as reusing the LLM's own latent space and propose the Invariant Latent Space Hypothesis (ILSH): (1) diverse outputs from the same source prompt should preserve consistent semantics (source invariance), and (2) input<->output cyclic mappings should be self-consistent within a shared latent space (cyclic invariance). Accordingly, we present Inv^2A, which treats the LLM as an invariant decoder and learns only a lightweight inverse encoder that maps outputs to a denoised pseudo-representation. When multiple outputs are available, they are sparsely concatenated at the representation layer to increase information density. Training proceeds in two stages: contrastive alignment (source invariance) and supervised reinforcement (cyclic invariance). An optional training-free neighborhood search can refine local performance. Across 9 datasets covering user and system prompt scenarios, Inv^2A outperforms baselines by an average of 4.77% BLEU score while reducing dependence on large inverse corpora. Our analysis further shows that prevalent defenses provide limited protection, underscoring the need for stronger strategies. The source code and data involved in this paper can be found in https://github.com/yyy01/Invariant_Attacker.
Open AccessAn Invariant Latent Space Perspective on Language Model Inversion
Language model inversion (LMI), i.e., recovering hidden prompts from outputs, emerges as a concrete threat to user privacy and system security. We recast LMI as reusing the LLM's own latent space and propose the Invariant Latent Space Hypothesis (ILSH): (1) diverse outputs from the same source prompt should preserve consistent semantics (source invariance), and (2) input<->output cyclic mappings should be self-consistent within a shared latent space (cyclic invariance). Accordingly, we present Inv^2A, which treats the LLM as an invariant decoder and learns only a lightweight inverse encoder that maps outputs to a denoised pseudo-representation. When multiple outputs are available, they are sparsely concatenated at the representation layer to increase information density. Training proceeds in two stages: contrastive alignment (source invariance) and supervised reinforcement (cyclic invariance). An optional training-free neighborhood search can refine local performance. Across 9 datasets covering user and system prompt scenarios, Inv^2A outperforms baselines by an average of 4.77% BLEU score while reducing dependence on large inverse corpora. Our analysis further shows that prevalent defenses provide limited protection, underscoring the need for stronger strategies. The source code and data involved in this paper can be found in https://github.com/yyy01/Invariant_Attacker.