Author: Ide-san

(From my LinkedIn article) I’m happy to announce that I just released the beta version of Generative Perturbation Analysis (GPA) on GitHub https://github.com/Idesan/gpa along with the paper, poster, and presentation slides to be presented at KDD 2023 next week.  GPA is an algorithm for anomaly attribution, whose task is to provide an explanation for an observed deviation in the black-box regression setting. It falls into the category of XAI (explainable AI). For comparison purposes, the repo provides a carefully designed Python implementation of major black- and white-box attribution algorithms: Hope it is useful to researchers and practitioners.

Happy to share that one of my latest papers on explainable AI (XAI), entitled “Generative Perturbation Analysis for Probabilistic Black-Box Anomaly Attribution,” has been accepted to KDD 2023 (The 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining), one of the top machine learning (ML) conferences. For more details, see my LinkedIn post.

Shared my thoughts on decentralized machine learning in LinkedIn.

Shared one of my favorite project at IBM in LinkedIn.

I wrote about my thoughts on “Explainable AI” in LinkedIn.

I happily announce that my new book titled “大規模計算時代の統計推論” has just come out. It’s available at Amazon.co.jp. The book is a Japanese translation of the celebrated book “Computer Age Statistical Inference” by maestro statisticians Prof. B. Efron and Prof. Hastie. I supervised the translation project and translated a few chapters myself. Although the topics covered are mostly basic, the description of those topics is deep and profound. What I was particularly interested in is the way statisticians look at machine learning. The authors say: In the absence of optimality criteria, either frequentist or Bayesian, the prediction community grades algorithmic excellence on performance within a catalog of often-visited examples such as…

A paper on decentralized collaborative learning has been accepted to IJCAI 19, one of the top AI conferences, for presentation at Macau, China, in August. The paper starts with giving a compliment to Bitcoin for its innovative idea for securing data consistency, which is usually called “Proof-of-Work”. Although the original paper describes the idea as voting by computation power, I think the key idea is not just a voting but its stochastic nature. My paper presents a simple but efficient way of doing “(relatively) secure” collaborative learning with a particular focus on IoT (internet-of-things) applications, where cryptographically rigorous security would be too much. Hope I can see many people sharing…

It’s been a while since I posted last time. Now that we are looking at the year-end holiday season, I’m giving here some updates for the year of 2018. Blockchain I was involved in a research project on Blockchain this year. I was curious what kind of people were working on it. Unfortunately, I didn’t get many collaboration opportunities, but I was able to finish two papers on a new framework of Blockchain (one has been accepted as a workshop paper, the other is under review). Probabilistic tensor regression model My paper on probabilistic tensor regression has been accepted to AAAI-19. A major application I described in the paper is…

My paper on an image-based traffic estimation method has just been published. Let me cite a part of my email sent to the project members and managers: It’s been more than five years since we started this project in Japan. When Katsuki-san, a new hire at TRL at that time, started looking at traffic images at accessKenya.com, no one imagined that this project would receive global attention in a year or so. It was way before IBM started the extensive campaign on Cognitive Computing. I’m glad to see that this project has become one of the major activities at Africa Lab. It was one of the most exciting projects in…

My paper on questionnaire data analysis has just come out. This paper is the first paper to propose an supervised extension of the Item Response Theory, which is an interesting probabilistic model developed in the area of psychometrics. Tsuyoshi Idé and Amit Dhurandhar, “Supervised item response models for informative prediction,” Knowledge and Information Systems, 51 (2017) 235-257. @Article{Ide2017KAIS, author=”Tsuyoshi Id{\’e} and Amit Dhurandhar”, title=”Supervised item response models for informative prediction”, journal=”Knowledge and Information Systems”, year=”2017″, volume=”51″, number=”1″, pages=”235–257″, }