SELECTED ACADEMIC CV

Hang Gu 顾航

Reliable multimodal learning under limited labels.

Lecturer & Teaching Section Head · 江苏财经职业技术学院

Updated 15 October 2026. Manuscript and project statuses are recorded as of September 2026.

Education

Universiti Malaya

MSc, Computer Science (Data Science)

Thesis: Dental Deformities Prediction Using Bayesian Networks among Adolescents

Huaiyin Normal University

Bachelor's degree, Computer Science and Technology

Research record

First author · Journal resubmission

Frozen CLIP Features for Low-Resource Multimodal Fake News Detection: Simple Adaptation, Token-Level Fusion and Data Artifacts

Resubmission to International Journal of Machine Learning and Cybernetics. Studies frozen CLIP adaptation through token-level co-attention, prototype classification, baseline comparisons, modality ablations and near-duplicate / label-association audits.

Related research code ↗

Submitted · DASFAA 2027

Validation Sensitivity in Limited-Label News Classification

Compares frozen CLIP class prompts and logistic-regression probes using five training seeds, 20 shared validation draws and 100 additional draws. Separates the effects of validation composition, scoring weights and epoch selection.

First author · Accepted, ISCAIT 2026

Design and Implementation of a Vision-Language Multimodal Behavior Prediction Model

Combines VideoMAE and DistilBERT with bidirectional cross-attention, gated residuals, text dropout and LSTM. Studies behavior classification and trajectory trends through implementation and ablation analysis.

Revised manuscript returned · Awaiting final review

融合 LSTM 与特征优化的系统故障智能预测方法研究

《计算机工程与应用》

Time- and frequency-domain features, random-forest selection / PCA, a two-layer LSTM and dynamic alert thresholds for multivariate fault prediction.

Accepted

机器学习在游戏角色行为建模中的应用探析

《网络游戏技术与应用》

A review of supervised learning, reinforcement learning and deep learning for behavior recognition, policy learning and dynamic decisions.

Master's thesis · 2023

Dental Deformities Prediction Using Bayesian Networks among Adolescents

Clinical-data cleaning and discretization in R, MMHC structure learning and bootstrap consensus networks; comparison of static and dynamic Bayesian networks for interpretable conditional-dependency analysis.

Led projects

Intelligent information retrieval using the OpenAI API

In progress

Huai'an natural science project · Project lead

Knowledge-base and system prototypes; document chunking, embeddings, retrieval and reranking, with attention to retrieval quality, response time and source traceability.

AI contract management system

Industry collaboration · Project lead · In progress

Document parsing, entity extraction, risk-clause identification and similar-clause retrieval; prototype implementation and internal evaluation.

Modular administration dashboard

Industry collaboration · Project lead · Completed

Reusable interface components and API conventions; requirements, development, integration, deployment and acceptance.

Generative AI for integrated learning, practice and assessment

Jiangsu higher-education teaching-reform project · Project lead · In progress

Proposed study of differentiated practice, immediate feedback and learning-behavior assessment in introductory computing courses.

Experience

江苏财经职业技术学院

Lecturer & Teaching Section Head

Computing teaching, curriculum revision, Hadoop/Spark practical platforms and coordination of an educational AI research team.

Whale Cloud Technology · Frontend Lead

Core frontend architecture for U Mobile onboarding; configurable forms and state machines. Work orders, monitoring, multilingual interfaces and weak-network support for an India water project.

Nanjing Siyue Network · Frontend Engineer

Printing ERP: orders, Gantt scheduling, progress tracking and mobile scan feedback.

Wangdao Technology · Frontend Engineer

Online learning and practical-training interfaces: courses, assignments, editing and automated assessment.

Public artifacts

  • KICP — multimodal research implementation and resource requirements.
  • KAPT — KANO-derived learning tools: early stopping, aligned ensembles and synthetic tests.
  • Counselor workflow case — role scopes, state transitions and release verification.
  • Frontend & Data Lab — educational examples of data preparation, annotation and evaluation.

Selected teaching contributions

  • 2025: coached a national first-prize team in the BRICS skills competition's blockchain application development event; received an excellent-instructor award and served as a national-event judge.
  • 2024 and 2025: coached second-prize teams in Jiangsu vocational-college big-data application development competitions.
  • 副主编《大数据技术在财务会计中的应用(Hive分析)》《人工智能基础与实践应用》;参编《智能制造应用技术与实践(第二版)》。
Read the proposed research direction ↗Inspect code and verification ↗