Qian Shao
Sentence-Level Evidence Extraction for Rumor Detection
Experience
Oct 2025 - Present
4 monthsSentence-Level Evidence Extraction for Rumor Detection
- Proposing a two-stage framework that identifies key evidence sentences from articles before rumor classification
- Exploring alternative strategies for sentence-level evidence extraction, including potential LLM-assisted approaches
- Designing interpretable architectures that connect evidence extraction with veracity prediction
- Currently focusing on problem formulation and framework design
Aug 2025 - Oct 2025
3 monthsSingapore
AI Intern
Doxa Holdings
- Designed and implemented the Buyer AI Agent for procurement analytics, combining rule-based logic with LLM for natural-language query reasoning
- Built end-to-end data pipelines covering purchase orders, requisitions, and delivery data to support intelligent buyer queries
Sep 2024 - Apr 2025
8 monthsSingapore
Research Intern
A*STAR, Institute for Infocomm Research
- Applied imitation learning to model expert treatment decision behaviors in healthcare
- Investigated inverse reinforcement learning (IRL) approaches for aligning AI policy learning with clinical reasoning
Jun 2024 - Oct 2024
5 monthsHybrid
LLM-based Few-Shot Early Rumor Detection with Imitation Agent
- Proposed a hybrid framework combining a lightweight imitation agent with a frozen LLM for early rumor detection
- Modeled the task as an MDP where the agent learns early-stopping policies from expert trajectories
- Integrated multi-expert imitation to balance earliness, stability, and accuracy
- improved macro-F1 and early detection rate across four benchmark datasets
Jan 2022 - May 2024
2 years 5 monthsConstrained Pricing in Choice-Based Revenue Management
- Studied dynamic pricing under multinomial logit (MNL) demand with capacity and budget constraints
- Developed a piecewise-linear approximation framework to address non-convex pricing objectives
- Modeled multi-product pricing decisions with revenue and inventory constraints in simulation
- Demonstrated superior solution quality and computational efficiency over gradient-based methods
Jan 2022 - Aug 2023
1 year 8 monthsImitating Cost-Constrained Behaviors in Reinforcement Learning
- Formulated the cost-constrained imitation learning problem where expert behaviors follow latent cost limits
- Proposed three algorithms—CCIL, MALM, and CVAG—combining Lagrangian, meta-gradient, and alternating updates
- Modeled expert trajectories using Safety-Gym and MuJoCo simulation environments
- Achieved the best balance between reward recovery and cost compliance across baselines
Apr 2021 - Dec 2021
9 monthsPreference-Aware Delivery Planning for Last-Mile Logistics
- Proposed a two-level route optimizer that integrates optimization and learnable driver preferences
- Modeled zone-level TSP and stop-level routing with Bayesian weight adjustment for feature relevance
- Implemented on Amazon Last Mile Challenge data to replicate practitioner-preferred delivery patterns
- Demonstrated improved route alignment and lower deviation from driver choices
Summary
AI researcher with PhD-level expertise in reinforcement learning, imitation learning, and optimization for decision-making under constraints. Experienced in developing agent-based and data-efficient learning frameworks across logistics, pricing, and rumor detection. Recent work explores LLM-based early rumor detection and intelligent agent design for real-world applications.
Languages
English
AdvancedChinese
AdvancedEducation
Aug 2020 - Jun 2025
Singapore Management University (SMU)
Ph.D., Reinforcement Learning, Imitation Learning, AI-Agent · Computer Science · Singapore · 3.53/4.0
Sep 2018 - Jun 2020
Jilin University
M.S. · Management Science and Engineeing · Changchun, China · 80.5/100
Sep 2013 - Jun 2017
Nanjing Agricultural University
B.S. · Information Management and Information System · Nanjing, China · 3.79/4.5
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience