研究 (Research)

Journal articles:

• P.-S. Chen, H.-Y. Liu, and C.-H. Hsu, Economic Order Quantity of Deteriorating Drugs with New and Redistributed Demands in a Closed-loop Supply Chain: From the Viewpoint of a Hospital Pharmacy, International Journal of Systems Science: Operations & Logistics, 2026, Vol. 13, No. 1. (SCIE) 
• P.-S. Chen, J. Lyu,  C.-W. Chen, I.-H. Yu, and C.-H. Hsu, The Social-economic Value of a Renewable Energy Power Generation and Energy Storage System: A Case Study, International Journal of Systems Science: Operations & Logistics, 2026, Vol. 13, No. 1. (SCIE)
• H.-A. Bui, C.-H. Hsu, H.-W.V. Young, Y.-Y. Chen, and Y.-A. Liou, Advanced Semi-Supervised Learning for Remote Sensing-Based Land Cover Classification in the Mekong River Delta, Vietnam, Remote Sensing, 2026, 18, 989. https://doi.org/10.3390/rs18070989. (SCIE)

The most-cited papers of the twenty-first century

Helen Pearson, Heidi Ledford, Matthew Hutson & Richard Van Noorden, Exclusive: the most-cited papers of the twenty-first century: A Nature analysis reveals the 25 highest-cited papers published this century and explores why they are breaking records, Nature, 15 April 2025. (Supplementary information)

Papers in AI:

Data-Driven Science and Engineering

Steven L. Brunton and J. Nathan Kutz, Data-Driven Science and Engineering: Machine Learning, Dynamical Systems, and Control, 2nd edition, Cambridge University Press (pdf, amazing videos, connection with dynamical systems in engineering)

Foundations of Computer Vision

Antonio Torralba, Phillip Isola, and William Freeman, Foundations of Computer Vision, The MIT Press, 2024.

作者是三位 MIT 的教授。如果你想了解電腦視覺 (Computer vision) 相關研究，甚至機器學習、研究方法，大力推荐。我有空就翻一下，增加對整個領域的了解，每次都有驚喜，很神奇的書。例如

Think Like an Amateur, Do As an Expert

作者：金出武雄，譯者：鄭舜瓏，像外行一樣思考，像專家一樣實踐：成功解決問題的高階技術，遠流

Jitendra S. Mudhol, Think Like an Amateur, Do As an Expert: Lessons from a Career in Computer Vision, CollaMeta, May 17, 2019. (video)

Serve Robotics Level 4 delivery

Serve Robotics: 新聞

Google DeepMind 科學家公布具備人類玩家業餘水準的桌球機器人

陳曉莉，DeepMind的桌球機器人已達業餘的人類選手水準，iThome，2024-08-09 

Data-Driven Performance Guarantees for Classical and Learned Optimizers

R. Sambharya and B. Stellato, Data-Driven Performance Guarantees for Classical and Learned Optimizers, arXiv e-prints:2404.13831,2024. (Python code)

AI 養魚解決老師傅技術失傳問題

邱倢芯，從魚缸到金目鱸養殖場，AI 養魚解決老師傅技術失傳問題，科技新報，2023 年 11 月 09 日

對於建置一套系統，許多業主最擔心的莫過於建置成本；對此，劉建伸坦言，系統本身的確不便宜，從導入初期至今已經投入 8 位數的成本，但後續帶來的效益也相當明顯，像是養殖戶過去都得長時間留守漁塭，在導入系統後，可將平均每天 8 小時的工作時間降低至 6.5 小時。

人力成本也可進一步降低，王靜儀估算，透過導入系統，每 100 公頃養殖面積的人力可從 33 人降低至 10 人，節省近 2,000 萬的人力成本。

智慧農業的成功因素

林一平，智慧農業的成功因素，電子時報，2023-08-21

在台灣，農業物聯網感測設備供應商眾多，然而通訊技術和資料傳輸格式卻千差萬別，導致資料在不同系統間的流通和加值應用面臨著困難。

為了解決這一重要問題，農業部於2023年4月27日推出「智慧農業感測資料格式標準與測試規範」。透過推動資料格式的標準化，提高農業物聯網應用領域中資料串接的效率，同時也降低開發成本，推動農業物聯網的深入應用。

Automated Machine Learning

Haifeng Jin, François Chollet, Qingquan Song, and Xia Hu. "AutoKeras: An AutoML Library for Deep Learning." the Journal of Machine Learning Research 6 (2023): 1-6. (Download)

To use deep learning, one needs to be familiar with various software tools like TensorFlow or Keras, as well as various model architecture and optimization best practices. Despite recent progress in software usability, deep learning remains a highly specialized occupation. To enable people with limited machine learning and programming experience to adopt deep learning, we developed AutoKeras, an Automated Machine Learning (AutoML) library that automates the process of model selection and hyperparameter tuning. AutoKeras encapsulates the complex process of building and training deep neural networks into a very simple and accessible interface, which enables novice users to solve standard machine learning problems with a few lines of code. Designed with practical applications in mind, AutoKeras is built on top of Keras and TensorFlow, and all AutoKeras-created models can be easily exported and deployed with the help of the TensorFlow ecosystem tooling.

Integration of Face-to-Face Screening With Real-time Machine Learning to Predict Risk of Suicide Among Adults

Drew Wilimitis, Robert W. Turer, Michael Ripperger, et al., Integration of Face-to-Face Screening With Real-time Machine Learning to Predict Risk of Suicide Among Adults, JAMA Netw Open. 2022; 5(5):e2212095. doi:10.1001/jamanetworkopen.2022.12095.

In this cohort study of 120 398 adult patient encounters, an ensemble learning approach combined suicide risk predictions from the Columbia Suicide Severity Rating Scale and a real-time machine learning model. Combined models outperformed either model alone for risks of suicide attempt and suicidal ideation across a variety of time periods.

Outracing champion Gran Turismo drivers with deep reinforcement learning

Wurman, P.R., Barrett, S., Kawamoto, K. et al. Outracing champion Gran Turismo drivers with deep reinforcement learning. Nature 602, 223–228 (2022). https://doi.org/10.1038/s41586-021-04357-7.

Many potential applications of artificial intelligence involve making real-time decisions in physical systems while interacting with humans. Automobile racing represents an extreme example of these conditions; drivers must execute complex tactical manoeuvres to pass or block opponents while operating their vehicles at their traction limits. Racing simulations, such as the PlayStation game Gran Turismo, faithfully reproduce the non-linear control challenges of real race cars while also encapsulating the complex multi-agent interactions. Here we describe how we trained agents for Gran Turismo that can compete with the world’s best e-sports drivers. We combine state-of-the-art, model-free, deep reinforcement learning algorithms with mixed-scenario training to learn an integrated control policy that combines exceptional speed with impressive tactics. In addition, we construct a reward function that enables the agent to be competitive while adhering to racing’s important, but under-specified, sportsmanship rules. We demonstrate the capabilities of our agent, Gran Turismo Sophy, by winning a head-to-head competition against four of the world’s best Gran Turismo drivers. By describing how we trained championship-level racers, we demonstrate the possibilities and challenges of using these techniques to control complex dynamical systems in domains where agents must respect imprecisely defined human norms.

Efficient method for training deep networks with unitary matrices

Bobak Kiani, Randall Balestriero, Yann Lecun, and Seth Lloyd,  projUNN: efficient method for training deep networks with unitary matrices, arXiv:2203.05483v2.

In learning with recurrent or very deep feed-forward networks, employing unitary matrices in each layer can be very effective at maintaining long-range stability. However, restricting network parameters to be unitary typically comes at the cost of expensive parameterizations or increased training runtime. We propose instead an efficient method based on rank-k updates -- or their rank-k approximation -- that maintains performance at a nearly optimal training runtime. We introduce two variants of this method, named Direct (projUNN-D) and Tangent (projUNN-T) projected Unitary Neural Networks, that can parameterize full N-dimensional unitary or orthogonal matrices with a training runtime scaling as O(kN^2). Our method either projects low-rank gradients onto the closest unitary matrix (projUNN-T) or transports unitary matrices in the direction of the low-rank gradient (projUNN-D). Even in the fastest setting (k=1), projUNN is able to train a model's unitary parameters to reach comparable performances against baseline implementations. By integrating our projUNN algorithm into both recurrent and convolutional neural networks, our models can closely match or exceed benchmarked results from state-of-the-art algorithms.

Strong mixed-integer programming formulations for trained neural networks

R. Anderson, J. Huchette, W. Ma, C. Tjandraatmadja, and J.P. Vielma, Strong mixed-integer programming formulations for trained neural networks, Mathematical Programming, 2020, 183(1-2):3-39, ISSN 14364646, URL http://dx.doi.org/10.1007/s10107-020-01474-5.

We present strong mixed-integer programming (MIP) formulations for high-dimensional piecewise linear functions that correspond to trained neural networks. These formulations can be used for a number of important tasks, such as verifying that an image classification network is robust to adversarial inputs, or solving decision problems where the objective function is a machine learning model. We present a generic framework, which may be of independent interest, that provides a way to construct sharp or ideal formulations for the maximum of d affine functions over arbitrary polyhedral input domains. We apply this result to derive MIP formulations for a number of the most popular nonlinear operations (e.g. ReLU and max pooling) that are strictly stronger than other approaches from the literature. We corroborate this computationally, showing that our formulations are able to offer substantial improvements in solve time on verification tasks for image classification networks.

Google Research: Themes from 2021 and Beyond

Jeff Dean, Google Research: Themes from 2021 and Beyond, Google, January 11, 2022.

· Trend 1: More Capable, General-Purpose ML Models

  · Trend 2: Continued Efficiency Improvements for ML

  · Trend 3: ML Is Becoming More Personally and Communally Beneficial

  · Trend 4: Growing Benefits of ML in Science, Health and Sustainability

  · Trend 5: Deeper and Broader Understanding of ML

YouTube video streaming now using A.I. that mastered chess and Go

JEREMY KAHN, YouTube video streaming now using A.I. that mastered chess and Go, Fortune, February 11, 2022.

The artificial intelligence algorithm, called MuZero, was developed by YouTube’s London-based sister company within Alphabet, DeepMind, which is dedicated to advanced A.I. research. When applied to YouTube videos, the system has resulted in a 4% reduction on average in the amount of data the video-sharing service needs to stream to users, with no noticeable loss in video quality.

自駕車挑戰賽

非凡財經新聞，比職業賽車手厲害?"演算法"當駕駛員 自駕車時速飆218km 德國團隊奪冠抱走2800萬獎金，20211025 (Indy Autonomous Challenge)

DeepMind’s AI predicts almost exactly when and where it’s going to rain

Will Douglas Heaven, DeepMind’s AI predicts almost exactly when and where it’s going to rain, MIT Technology Review, September 29, 2021.

In a blind comparison with existing tools, several dozen experts judged DGMR’s forecasts to be the best across a range of factors—including its predictions of the location, extent, movement, and intensity of the rain—89% of the time. The results were published in a Nature paper today.

科技部半導體射月計畫 109 年度產學技術交流

 科技部半導體射月計畫 109 年度產學技術交流

• Boris Murmann, tinyML: The Perfect Storm for Innovation in Ultra-Low-Power System Design 
• 梁伯嵩，IC 運算平台趨勢: 數位運算、人工智慧與量子運算
• Tetsu Ohtou, Semiconductor Process and Equipment Technology for Advanced Logic Devices 
• 陳俊雄，汽車產業及感測元件發展趨勢