J.-L. Li | Po-Chung (Bryan) Hsiehhttps://pchsieh.com/authors/j.-l.-li/J.-L. LiHugoBlox Kit (https://hugoblox.com)en-usSat, 21 Mar 2026 00:00:00 +0000The Binding Effect: Analyzing How Multi-Dimensional Cues Form Gender Bias in Instruction TTShttps://pchsieh.com/publications/binding-effect-gender-bias-instruction-tts/Sat, 21 Mar 2026 00:00:00 +0000https://pchsieh.com/publications/binding-effect-gender-bias-instruction-tts/<p>This paper analyzes how multi-dimensional cues in natural-language instructions bind together to form gender bias in instruction-following text-to-speech (TTS) systems.</p> <p>Accepted to <strong>INTERSPEECH 2026</strong>.</p>Dual-Process Scaffold Reasoning for Enhancing LLM Code Debugginghttps://pchsieh.com/publications/dual-process-scaffold-reasoning-llm-code-debugging/Tue, 11 Nov 2025 00:00:00 +0000https://pchsieh.com/publications/dual-process-scaffold-reasoning-llm-code-debugging/<p>This paper proposes a dual-process scaffold reasoning framework for large language model (LLM) code debugging. It reaches an <strong>88.91% pass rate on DebugBench</strong>, outperforming prior prompting methods.</p> <p><strong>My role:</strong> first author. This work was done during my LLM engineer internship at the Inventec Artificial Intelligence Center.</p>