Published in , 2025
Cascaded dual-encoder retrieval plus ontology reasoning to normalize adverse drug events to MedDRA terms.
Recommended citation: Dao-Sy, D.-M., Nguyen-Lam, P.-Q., Pham, P.-H., Tran, C.-N., Huynh, T.-K., Truong, B.-T. (2025). DRAGON: Dual-Encoder Retrieval with Guided Ontology Reasoning for Medical Normalization. ALTA 2025, pp. 230-239.
Download Paper
Published in , 2025
Classical ML against transformer models for classifying telescope-related scientific literature.
Recommended citation: Huynh, T.-K., Dao-Sy, D.-M., Tran, C.-N., Nguyen-Lam, P.-Q., Pham, P.-H., Nguyen, D.-H.-D., Dinh, D., Nguyen, H.-B.-L. (2025). Systematic Evaluation of Machine Learning and Transformer-Based Methods for Scientific Telescope Literature Classification. WASP @ IJCNLP-AACL 2025, pp. 136-145.
Download Paper
Published in , 2025
Retrieval, LoRA-tuned NLLB and a lexical SMT branch, combined by Minimum Bayes Risk decoding for low-resource Indic translation.
Recommended citation: Huynh, T.-K., Dao-Sy, D.-M., Tran, C.-N., Nguyen-Lam, P.-Q., Pham, P.-H., Nguyen, D.-H.-D., Dinh, D., Nguyen, H.-B.-L. (2025). JHARNA-MT. MMLoSo 2025, pp. 114-120.
Download Paper
Published in , 2026
Predicts at creation time which agent-authored pull requests will cost maintainers the most review effort; flagging the top 20% catches 69% of total effort.
Recommended citation: Dao-Sy, D.-M., Huynh, T.-K., Nguyen-Lam, P.-Q., Pham, P.-H., Tran, C.-N., Nguyen, D.-H.-D., Truong, B.-T. (2026). Early-Stage Prediction of Review Effort in AI-Generated Pull Requests. MSR 2026 Mining Challenge. arXiv:2601.00753.
Published in , 2026
How the size of the stake and the prompt language change the strategies LLM agents play in cooperation dilemmas. Earlier title: More at Stake.
Recommended citation: Huynh, T.-K., Dao-Sy, D.-M., Cao, T.-B., et al., Tran, C.-N., et al., Han, T. A. (2026). Payoff Scaling Shapes Cooperation in LLM Agents across Languages. Interface Focus (accepted).
Published in , 2026
Transformer-based test prioritization for self-driving-car simulation testing, APFD 0.804.
Recommended citation: Tran, C.-N., Dao-Sy, D.-M., Huynh, T.-K., Pham, P.-H., Nguyen-Lam, P.-Q. (2026). RoadFury at the ICST 2026 Tool Competition: Self-Driving Car Testing Track. IEEE ICST 2026, pp. 509-510.
Download Paper
Published in , 2026
The benchmark paper of the NeurIPS 2025 Mouse vs. AI competition, co-authored with the organizers as a member of the first-place team.
Recommended citation: Schneider, M., Canzano, J. S., Hou, Y., Peng, J., Deepu, A., Karan, U., Pham, P.-H., Tran, C.-N., Dao-Sy, D.-M., Nguyen-Lam, P.-Q., Huynh, T.-K., Azeglio, S., Smith, S., Beyeler, M. (2026). Visual Robustness and Neural Alignment in a Shared Foraging Task: The Mouse vs. AI Benchmark. NeurIPS 2026 Evaluations and Datasets Track.
Published in , 2026
Aligns a frozen LLM to the moral preferences of a country without training: country-conditioned personas vote, and only the decisions they agree on are kept.
Recommended citation: Huynh, T.-K., Dao-Sy, D.-M., Nguyen, T., Tran, C.-N., Pham, P.-H., Nguyen-Lam, P.-Q., Han, T. A., Tran-Thanh, L. (2026). Training-Free Cultural Alignment of Large Language Models via Persona Disagreement. NeurIPS 2026. arXiv:2605.10843.
Published in , 2026
Holistic training with a slot-attention router, then compositional inference that scores queries part by part, so a continual few-shot learner generalizes to unseen combinations of known parts.
Recommended citation: Nguyen-Lam, P.-Q., Pham, P.-H., Dao-Sy, D.-M., Tran, C.-N., Huynh, T.-K., Tran-Thanh, L. (2026). Unlocking Compositional Generalization in Continual Few-Shot Learning. NeurIPS 2026. arXiv:2605.11710.
Published in , 2026
Streaming decision trees that learn new classes online from a data stream, with split decisions backed by a McDiarmid concentration bound and knowledge inherited across splits.
Recommended citation: Pham, P.-H.*, Tran, C.-N.*, Nguyen-Lam, P.-Q., Dao-Sy, D.-M., Huynh, T.-K., Tran-Thanh, L. (2026). MIST: Reliable Streaming Decision Trees for Online Class-Incremental Learning via McDiarmid Bound. NeurIPS 2026. arXiv:2605.11617.
Published in , 2026
SkyPart: a lightweight prototype-based semantic-part head on a DINOv2 ViT that matches drone views to satellite tiles under fog, rain, snow and darkness. State of the art on weather-robust drone-to-satellite geo-localization with the fewest parameters among top methods.
Recommended citation: Tran, C.-N.*, Dao-Sy, D.-M.*, Huynh, T.-K.*, Nguyen-Lam, P.-Q., Pham, P.-H., Tran-Thanh, L. (2026). Weather-Robust Cross-View Geo-Localization via Prototype-Based Semantic Part Discovery. NeurIPS 2026. arXiv:2605.11654.
Published in , 2026
An agentic resolver that repairs broken Python environments with a self-evolving memory, a confidence cascade over version candidates, and a build-and-reflect loop.
Recommended citation: Tran, C.-N., Dao-Sy, D.-M., Huynh, T.-K., Pham, P.-H., Nguyen-Lam, P.-Q., Nguyen, V. (2026). MEMRES: A Memory-Augmented Resolver with Confidence Cascade for Agentic Python Dependency Resolution. FSE Companion 2026, pp. 1438-1441.
Download Paper
Published in , 2026
Finds the most plausible cause of a real event from retrieved documents with hybrid retrieval and extended LoRA.
Recommended citation: Dao-Sy, D.-M.*, Tran, C.-N.*, Huynh, T.-K.*, Nguyen-Lam, P.-Q., Pham, P.-H. (2026). HCMUS_RepeatedGames at SemEval-2026 Task 12: CausalRAG: Synergizing Causal Graph Retrieval and Extended LoRA for Abductive Reasoning. SemEval-2026, pp. 409-416.
Download Paper
Published in , 2026
Separates logical form from content so an LLM judges whether a syllogism is valid, not whether it sounds plausible.
Recommended citation: Tran, C.-N.*, Dao-Sy, D.-M.*, Huynh, T.-K., Pham, P.-H., Nguyen-Lam, P.-Q. (2026). HCMUSDroneBoys at SemEval-2026 Task 11: Asymmetric Counterfactual Debiasing and Rank-Sensitive Logical Invariance Adaptation for Syllogistic Reasoning. SemEval-2026, pp. 658-663.
Download Paper