Selected work

Representative Papers

Demystifying Chains, Trees, and Graphs of Thoughts

Jan 25, 2024

Existing structured prompting paradigms—such as Chain-of-Thought (CoT), Tree-of-Thought (ToT), and Graph-of-Thought (GoT)—lack a unified theoretical foundation, suffering from conceptual conflation and an absence of systematic taxonomy. Method: We propose the first comprehensive taxonomy for structured prompting, formally defining the notion of “reasoning topology,” constructing its spatial representation, and unifying CoT, ToT, and GoT through pipeline-based execution analysis, structural modeling, behavioral interpretation, and cross-paradigm empirical comparison. Contribution/Results: (1) We establish the first principled taxonomy for structured-prompt reasoning; (2) we uncover intrinsic relationships between topological structure and both reasoning performance and computational cost; and (3) we provide a theoretically grounded framework and design principles for scalable, interpretable prompt engineering.

21 citations1 influentialRead paper

Quality Without Usefulness: LLM-Generated XAI Narratives as Trust Heuristics Rather Than Decision Aids

May 26, 2026

This study investigates whether high-quality natural language explanations generated by large language models genuinely enhance users’ decision-making performance in real-world tasks or merely induce illusory trust. Through five controlled experiments in a time-series energy forecasting setting, the authors systematically evaluate the impact of such explanations on task accuracy and reliability judgments, employing factorial designs, placebo controls, and out-of-distribution detection. They identify and formally define a “quality-usefulness gap” in explainable AI (XAI): explanations exhibiting high textual quality do not necessarily support effective decision-making; instead, their mere presence inflates user confidence and impairs the ability to detect model failures. Results demonstrate that these explanations fail to improve accuracy across five task types and can even lead to misleading overconfidence.

0 citationsRead paper

Meta-learning for sample-efficient Bayesian optimisation of fed-batch processes

May 06, 2026

This work addresses the limitations of conventional Gaussian processes in Bayesian optimization, which struggle to model unobservable inter-batch variations, exhibit poor generalization, and suffer from low data efficiency. To overcome these challenges, the study introduces System-Aware Neural ODE Processes (SANODEP) into the Bayesian optimization framework for the first time, integrating meta-learning to construct a prior model capable of effectively capturing time-varying stochastic batch dynamics. The proposed approach substantially enhances both generalization performance and sample efficiency under both in-distribution and out-of-distribution batch conditions. Demonstrated on a penicillin batch production case study, the method achieves superior optimization outcomes with only a small number of experimental trials, yielding better objective values more rapidly than traditional Gaussian process-based approaches and significantly accelerating the initial optimization phase.

0 citationsRead paper

Fractal Characterization of Low-Correlation Signals in AI-Generated Image Detection

Apr 19, 2026

This work addresses the limited robustness of existing deepfake detection methods in open-world scenarios, where distinguishing highly realistic AI-generated images from authentic ones remains challenging. The authors propose a novel signal-level approach that, for the first time, integrates fractal theory into low-correlation signal analysis to uncover subtle statistical anomalies introduced during image synthesis by generative models. By operating directly on intrinsic signal characteristics rather than relying on high-level semantic features, the method overcomes key limitations of conventional detectors. The resulting general-purpose, signal-based detection model demonstrates superior robustness and cross-dataset generalization, significantly improving detection accuracy across diverse AI-generated imagery in unconstrained real-world settings.

0 citationsRead paper

From Uniform to Learned Knots: A Study of Spline-Based Numerical Encodings for Tabular Deep Learning

Apr 07, 2026

This study systematically investigates the impact of numerical feature preprocessing on tabular deep learning performance, with a focus on spline-based encoding for continuous variables. The authors propose a differentiable knot parameterization method that enables end-to-end joint optimization of knot locations for B-splines, M-splines, and I-splines alongside backbone architectures such as MLPs, ResNets, and FT-Transformers. They empirically compare uniform, quantile-based, target-aware, and learnable knot placement strategies across classification and regression tasks. Results show that piecewise linear encoding yields the most robust performance in classification settings, while regression accuracy is highly sensitive to the choice of spline type, knot strategy, and output dimensionality. Although learnable knots can enhance performance, they substantially increase training overhead.

0 citationsRead paper

Recent publications

Latest Papers

Quality Without Usefulness: LLM-Generated XAI Narratives as Trust Heuristics Rather Than Decision Aids

May 26, 2026

This study investigates whether high-quality natural language explanations generated by large language models genuinely enhance users’ decision-making performance in real-world tasks or merely induce illusory trust. Through five controlled experiments in a time-series energy forecasting setting, the authors systematically evaluate the impact of such explanations on task accuracy and reliability judgments, employing factorial designs, placebo controls, and out-of-distribution detection. They identify and formally define a “quality-usefulness gap” in explainable AI (XAI): explanations exhibiting high textual quality do not necessarily support effective decision-making; instead, their mere presence inflates user confidence and impairs the ability to detect model failures. Results demonstrate that these explanations fail to improve accuracy across five task types and can even lead to misleading overconfidence.

0 citationsRead paper

Meta-learning for sample-efficient Bayesian optimisation of fed-batch processes

May 06, 2026

This work addresses the limitations of conventional Gaussian processes in Bayesian optimization, which struggle to model unobservable inter-batch variations, exhibit poor generalization, and suffer from low data efficiency. To overcome these challenges, the study introduces System-Aware Neural ODE Processes (SANODEP) into the Bayesian optimization framework for the first time, integrating meta-learning to construct a prior model capable of effectively capturing time-varying stochastic batch dynamics. The proposed approach substantially enhances both generalization performance and sample efficiency under both in-distribution and out-of-distribution batch conditions. Demonstrated on a penicillin batch production case study, the method achieves superior optimization outcomes with only a small number of experimental trials, yielding better objective values more rapidly than traditional Gaussian process-based approaches and significantly accelerating the initial optimization phase.

0 citationsRead paper

Fractal Characterization of Low-Correlation Signals in AI-Generated Image Detection

Apr 19, 2026

This work addresses the limited robustness of existing deepfake detection methods in open-world scenarios, where distinguishing highly realistic AI-generated images from authentic ones remains challenging. The authors propose a novel signal-level approach that, for the first time, integrates fractal theory into low-correlation signal analysis to uncover subtle statistical anomalies introduced during image synthesis by generative models. By operating directly on intrinsic signal characteristics rather than relying on high-level semantic features, the method overcomes key limitations of conventional detectors. The resulting general-purpose, signal-based detection model demonstrates superior robustness and cross-dataset generalization, significantly improving detection accuracy across diverse AI-generated imagery in unconstrained real-world settings.

0 citationsRead paper

From Uniform to Learned Knots: A Study of Spline-Based Numerical Encodings for Tabular Deep Learning

Apr 07, 2026

This study systematically investigates the impact of numerical feature preprocessing on tabular deep learning performance, with a focus on spline-based encoding for continuous variables. The authors propose a differentiable knot parameterization method that enables end-to-end joint optimization of knot locations for B-splines, M-splines, and I-splines alongside backbone architectures such as MLPs, ResNets, and FT-Transformers. They empirically compare uniform, quantile-based, target-aware, and learnable knot placement strategies across classification and regression tasks. Results show that piecewise linear encoding yields the most robust performance in classification settings, while regression accuracy is highly sensitive to the choice of spline type, knot strategy, and output dimensionality. Although learnable knots can enhance performance, they substantially increase training overhead.

0 citationsRead paper

GraphWalk: Enabling Reasoning in Large Language Models through Tool-Based Graph Navigation

Apr 02, 2026

Current large language models are constrained by context length, making it challenging to perform multi-hop reasoning and complex queries over enterprise-scale knowledge graphs. This work proposes a training-free, task-agnostic tool-augmented framework that introduces, for the first time, a minimal orthogonal set of graph operations, enabling general-purpose large language models to traverse graph structures in a sequential, transparent, and verifiable manner to accomplish multi-step reasoning. By integrating tool-call-driven graph traversal, decomposition of multi-hop queries, and retrieval-augmented generation, the approach substantially outperforms in-context reasoning on both synthetic and enterprise-like knowledge graphs. The performance gain becomes more pronounced with larger model scales, effectively mitigating the performance collapse commonly observed in traditional methods when handling complex queries.

0 citationsRead paper