About

Chris Kuo is a seasoned data science professional with more than 25 years of experience applying advanced analytics across multiple industries. He has led high-impact data science initiatives in customer analytics, healthcare, fraud detection, and litigation support. He is also the inventor of a U.S. patent in data-driven solutions. Throughout his career, Chris has held leadership roles at several Fortune 500 companies in the insurance and retail sectors. He earned a BS in nuclear engineering from National Tsing Hua University in Taiwan and a PhD in economics from the State University of New York at Stony Brook. He has published research in economics and management journals and is the author of several books on data science, including The Handbook of Anomaly Detection (2024), Modern Time Series Forecasting Techniques for Predictive Analytics and Anomaly Detection (2024), and Introduction to Statistics with Python: The Essential Guide to Data Analysis (2025).

Research topics

• Machine Learning
• Artificial Intelligence
• Computer Science
• Computer Security
• Data science
• Computer vision

Selected publications

• Automatic Modulation Classification via Green Machine Learning

  arXiv (Cornell University) · 2026-04-11

  articleOpen accessSenior author

  In this work, we propose an interpretable, robust, and lightweight machine learning method for automatic modulation classification (AMC) under dynamic and noisy channel conditions. It is called green automatic modulation classification (GAMC) and targets edge artificial intelligence (AI) with low computational complexity and a small model size. GAMC operates in four stages. First, raw received I/Q signals are transformed into multi-domain representations, including constellation diagrams and spatio-temporal graphs. Second, we extract a comprehensive set of statistical and topological features from time-series signals, constellation diagrams, and graphs. Third, a supervised feature learning process leverages label guidance to project high-dimensional features into robust, discriminative low-dimensional ones. Finally, a context-aware Signal-to-Noise Ratio (SNR) soft routing mechanism ensembles predictions from downstream classifiers. Experimental results show that GAMC effectively mitigates domain shifts caused by high noise. It strikes a good balance between accuracy and efficiency, reducing the number of model parameters by $50\%$, operating at $3\%$ to $42\%$ of the computational cost of lightweight deep learning models, and maintaining higher accuracy in various SNRs.

  PublisherOA PDF

• Artificial intelligence-enabled flexible surface-enhanced Raman scattering substrate based on silver nanoparticles/polypyrrole/chitosan film for sensitive uric acid detection in saliva, serum and urine

  Carbohydrate Polymers · 2026-03-28

  article
  PublisherDOI

• Descrip3D: Enhancing Large Language Model-based 3D Scene Understanding with Object-Level Text Descriptions

  2026-03-06

  articleOpen accessSenior author

  Understanding 3D scenes goes beyond simply recognizing objects; it requires reasoning about the spatial and semantic relationships between them. Current 3D scenelanguage models often struggle with this relational understanding, particularly when visual embeddings alone do not adequately convey the roles and interactions of objects. In this paper, we introduce Descrip3D, a novel and powerful framework that explicitly encodes the relationships between objects using natural language. Unlike previous methods that rely only on 2D and 3D embeddings, Descrip3D enhances each object with a textual description that captures both its intrinsic attributes and contextual relationships. These relational cues are incorporated into the model through a dual-level integration: embedding fusion and prompt-level injection. This allows for unified reasoning across various tasks such as grounding, captioning, and question answering, all without the need for task-specific heads or additional supervision. When evaluated on five benchmark datasets, including ScanRefer, Multi3DRefer, ScanQA, SQA3D, and Scan2Cap, Descrip3D consistently outperforms strong baseline models, demonstrating the effectiveness of language-guided relational representation for understanding complex indoor scenes.

  PublisherOA PDFDOI

• Oxygen-driven electro-reduction of flutamide using a heterostructured Bi2S3–V2C ceramic nanocomposite: Toward sustainable environmental sensing applications

  Materials Chemistry and Physics · 2026-03-19

  articleCorresponding
  PublisherDOI

• Automatic Modulation Classification via Green Machine Learning

  arXiv (Cornell University) · 2026-04-11

  preprintOpen accessSenior author

  In this work, we propose an interpretable, robust, and lightweight machine learning method for automatic modulation classification (AMC) under dynamic and noisy channel conditions. It is called green automatic modulation classification (GAMC) and targets edge artificial intelligence (AI) with low computational complexity and a small model size. GAMC operates in four stages. First, raw received I/Q signals are transformed into multi-domain representations, including constellation diagrams and spatio-temporal graphs. Second, we extract a comprehensive set of statistical and topological features from time-series signals, constellation diagrams, and graphs. Third, a supervised feature learning process leverages label guidance to project high-dimensional features into robust, discriminative low-dimensional ones. Finally, a context-aware Signal-to-Noise Ratio (SNR) soft routing mechanism ensembles predictions from downstream classifiers. Experimental results show that GAMC effectively mitigates domain shifts caused by high noise. It strikes a good balance between accuracy and efficiency, reducing the number of model parameters by $50\%$, operating at $3\%$ to $42\%$ of the computational cost of lightweight deep learning models, and maintaining higher accuracy in various SNRs.

  PublisherDOI

• Phthalocyanine-derived dual-atom nickel catalysts (Ni-DAC) for efficient electrocatalytic CO₂ reduction

  Electrochimica Acta · 2026-03-10 · 1 citations

  articleCorresponding
  PublisherDOI

• Mass spectrometry-based metabolomics for studying gut microbiota-derived metabolites in biological samples

  Journal of Food and Drug Analysis · 2026-03-31

  articleOpen accessSenior author

  Gut microbiota produces a wide range of metabolites and plays a critical role in maintaining host health. Dysregulation of these metabolites can influence host metabolism through systemic circulation, contributing to the development of various diseases, including immunological, neurological and cancer-related disorders. Chromatography coupled with mass spectrometry (MS) has emerged as a powerful analytical approach, offering high sensitivity and resolution for studying gut microbiota-related metabolites. This review provides a comprehensive overview of chromatographic MS-based methods applied to the study of the gut microbial metabolome. We summarize strategies for sample collection, storage, and preparation of commonly analyzed sample types, including feces, plasma/serum, urine, and tissue samples. In addition, we included the main chromatographic MS-based approaches, as well as data analysis techniques, for investigating the gut microbial metabolome. The characteristics and utility of liquid chromatographic-mass spectrometry (LC-MS), gas chromatographic-mass spectrometry (GC-MS), and capillary electrophoresis-mass spectrometry (CE-MS) were discussed in the context of providing broader coverage of gut microbiota-derived metabolites with diverse physicochemical properties. Finally, we summarize recent studies that have employed chromatographic MS-based approaches to investigate gut microbiota-related disease. Through the integration of appropriate sample handling and advanced analytical strategies, a deeper understanding of host–microbiota interactions and their roles in health and disease can be achieved.

  PublisherDOI

• Cross-Modal Fine-Tuning of 3D Convolutional Foundation Models for ADHD Classification with Low-Rank Adaptation

  2026-04-08

  articleOpen access

  Early diagnosis of attention-deficit/hyperactivity disorder (ADHD) in children plays a crucial role in improving outcomes in education and mental health. Diagnosing ADHD using neuroimaging data, however, remains challenging due to heterogeneous presentations and overlapping symptoms with other conditions. To address this, we propose a novel parameter-efficient transfer learning approach that adapts a large-scale 3D convolutional foundation model, pre-trained on CT images, to an MRI-based ADHD classification task. Our method introduces Low-Rank Adaptation (LoRA) in 3D by factorizing 3D convolutional kernels into 2D low-rank updates, dramatically reducing trainable parameters while achieving superior performance. In a five-fold cross-validated evaluation on a public diffusion MRI database, our 3D LoRA fine-tuning strategy achieved state-of-the-art results, with one model variant reaching 71.9% accuracy and another attaining an AUC of 0.716. Both variants use only 1.64 million trainable parameters (over <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$113 \times$</tex> fewer than a fully fine-tuned foundation model). Our results represent one of the first successful cross-modal (CT-to-MRI) adaptations of a foundation model in neuroimaging, establishing a new benchmark for ADHD classification while greatly improving efficiency.

  PublisherOA PDFDOI

• GIFT: A Green Image Forgery Tracker

  Journal of Visual Communication and Image Representation · 2026-04-01

  articleSenior author
  PublisherDOI

• Low-Rate Denial-of-Service Attack Detection in Open Radio Access Network: Integrating Federated Learning

  IEEE Vehicular Technology Magazine · 2025-11-10

  article

  Security vulnerabilities have become increasingly critical with the growing connection of Internet of Things (IoT) devices and industrial control systems to 4G/5G private networks (PNs). Attacks in such environments are often complex and diverse, requiring advanced and adaptive detection mechanisms. However, conventional detection systems, which typically rely on the core network, incur high costs, introduce latency, and lack adaptability. The open radio access network (O-RAN) architecture provides a flexible and cost-effective framework by decoupling hardware and software components. This article presents a novel solution, the O-RAN advanced information security sharing system (O-RAN AIS3), which integrates federated learning (FL) to enhance detection capabilities. Detection models are deployed in near real-time RAN intelligent controllers (Near-RT RICs) in diverse deployment scenarios and are periodically updated through parameter sharing. A Y1 consumer aggregates parameters from multiple Near-RT RICs, enabling knowledge synchronization across diverse network environments. This collaborative learning approach enables cross-domain knowledge sharing, which is essential for addressing the dynamic nature of attacks in PNs and optimizing the detection of emerging threats in real time. We evaluated the effectiveness of this approach in detecting unseen low-rate and high-rate denial-of-service (DoS) attacks, leveraging FL for enhanced cross-operator security. The dataset, collected in an emulated PN environment, included traffic generated from 11 distinct attack types, covering both low-rate and high-rate scenarios. Experiments conducted in two 5G O-RAN PN environments demonstrated improved detection accuracy in both cases. These results underscore the effectiveness of FL-based systems in addressing the increasing complexity of malicious behaviors in operational PNs.

  PublisherDOI

Frequent coauthors

• Alex Acero

  Apple (Israel)

  896 shared

• Anna Scaglione

  Cornell University

  896 shared

• Mari Ostendorf

  896 shared

• Petar M. Djurić

  Stony Brook University

  889 shared

• J Treichler

  889 shared

• W.C. Karl

  Boston University

  889 shared

• Sergios Theodoridis

  National and Kapodistrian University of Athens

  889 shared

• Alex C. Kot

  852 shared

Education

• PhD, EECS

  Massachusetts Institute of Technology

  1987

• MS, EECS

  Massachusetts Institute of Technology

  1985