本スライドは、弊社の梅本により弊社内の技術勉強会で使用されたものです。
近年注目を集めるアーキテクチャーである「Transformer」の解説スライドとなっております。
"Arithmer Seminar" is weekly held, where professionals from within and outside our company give lectures on their respective expertise.
The slides are made by the lecturer from outside our company, and shared here with his/her permission.
Arithmer株式会社は東京大学大学院数理科学研究科発の数学の会社です。私達は現代数学を応用して、様々な分野のソリューションに、新しい高度AIシステムを導入しています。AIをいかに上手に使って仕事を効率化するか、そして人々の役に立つ結果を生み出すのか、それを考えるのが私たちの仕事です。
Arithmer began at the University of Tokyo Graduate School of Mathematical Sciences. Today, our research of modern mathematics and AI systems has the capability of providing solutions when dealing with tough complex issues. At Arithmer we believe it is our job to realize the functions of AI through improving work efficiency and producing more useful results for society.
Metric Recovery from Unweighted k-NN Graphsjoisino
Introduction of
- Towards Principled User-side Recommender Systems (CIKM 2022) https://arxiv.org/abs/2208.09864
- Graph Neural Networks can Recover the Hidden Features Solely from the Graph Structure (ICML 2023) https://arxiv.org/abs/2301.10956
- and their related technology.
Speakerdeck: https://speakerdeck.com/joisino/metric-recovery-from-unweighted-k-nn-graphs
本スライドは、弊社の梅本により弊社内の技術勉強会で使用されたものです。
近年注目を集めるアーキテクチャーである「Transformer」の解説スライドとなっております。
"Arithmer Seminar" is weekly held, where professionals from within and outside our company give lectures on their respective expertise.
The slides are made by the lecturer from outside our company, and shared here with his/her permission.
Arithmer株式会社は東京大学大学院数理科学研究科発の数学の会社です。私達は現代数学を応用して、様々な分野のソリューションに、新しい高度AIシステムを導入しています。AIをいかに上手に使って仕事を効率化するか、そして人々の役に立つ結果を生み出すのか、それを考えるのが私たちの仕事です。
Arithmer began at the University of Tokyo Graduate School of Mathematical Sciences. Today, our research of modern mathematics and AI systems has the capability of providing solutions when dealing with tough complex issues. At Arithmer we believe it is our job to realize the functions of AI through improving work efficiency and producing more useful results for society.
Metric Recovery from Unweighted k-NN Graphsjoisino
Introduction of
- Towards Principled User-side Recommender Systems (CIKM 2022) https://arxiv.org/abs/2208.09864
- Graph Neural Networks can Recover the Hidden Features Solely from the Graph Structure (ICML 2023) https://arxiv.org/abs/2301.10956
- and their related technology.
Speakerdeck: https://speakerdeck.com/joisino/metric-recovery-from-unweighted-k-nn-graphs
Towards Principled User-side Recommender Systemsjoisino
Ryoma Sato proposes a method called Consul for building user-side recommender systems when the system provided by a service like Twitter is unsatisfactory. Consul allows users to build recommender systems using only the information available to them through web pages, without having access to the full database. It does this while maintaining consistency with the official system, ensuring diversity in recommendations based on sensitive attributes, and being locally efficient without downloading all pages. Experiments show Consul performs as well as existing methods but is much more efficient due to its localized traversal of the recommendation graph. A case study demonstrates a user successfully building a new recommender system for Twitter using Consul.
CLEAR: A Fully User-side Image Search Systemjoisino
This document describes CLEAR, a fully user-side image search system developed by Ryoma Sato at Kyoto University. CLEAR allows users to build and publish their own image search engines without backend servers by formulating image search as a multi-armed bandit problem and implementing the system using only JavaScript on the client-side. This overcomes limitations of traditional search engines which require extensive resources to operate. CLEAR demonstrates that ordinary users can now develop customized image search tools.
Private Recommender Systems: How Can Users Build Their Own Fair Recommender S...joisino
JSSST 2022 https://jssst2022.wordpress.com/ における発表スライドです。
論文
Private Recommender Systems: How Can Users Build Their Own Fair Recommender Systems without Log Data? (SDM 2022)
arXiv: https://arxiv.org/abs/2105.12353
This document provides an introduction to spectral graph theory. It discusses how spectral graph theory connects combinatorics and algebra through studying graphs using eigenvalues and eigenvectors of adjacency matrices. It covers applications of spectral graph theory such as spectral clustering, which uses eigenvectors of the graph Laplacian as features for clustering nodes, and graph convolutional networks, which apply graph filtering and node-wise transformations to classify nodes in a graph.
第6回 統計・機械学習若手シンポジウムの公演で使用したユーザーサイド情報検索システムについてのスライドです。
https://sites.google.com/view/statsmlsymposium21/
Private Recommender Systems: How Can Users Build Their Own Fair Recommender Systems without Log Data? (SDM 2022) https://arxiv.org/abs/2105.12353
Retrieving Black-box Optimal Images from External Databases (WSDM 2022) https://arxiv.org/abs/2112.14921
Random Features Strengthen Graph Neural Networksjoisino
This document proposes using random features to strengthen graph neural networks (GNNs) for node classification tasks. It summarizes that GNNs cannot distinguish nodes with identical features and are not universal approximators. By adding random features to each node, GNNs can distinguish nodes and tree views, allowing them to detect graph structures like triangles. Experiments on synthetic and real-world graphs show random feature GNNs outperform standard GNNs and are a simple way to boost GNN expressiveness and performance.
51. 51 / 54 KYOTO UNIVERSITY
参考える文献:
Matti Åstrand, Patrik Floréen, Valentin Polishchuk, Joel Rybicki, Jukka Suomela,
Jara Uitto. A local 2-approximation algorithm for the vertex cover problem. DISC
2009.
Irwan Bello, Hieu Pham, Quoc V. Le, Mohammad Norouzi, Samy Bengio. Neural
Combinatorial Optimization with Reinforcement Learning. arXiv 2016.
Zhengdao Chen, Soledad Villar, Lei Chen, Joan Bruna. On the equivalence
between graph isomorphism testing and function approximation with GNNs.
NeurIPS 2019.
Hanjun Dai, Elias B. Khalil, Yuyu Zhang, Bistra Dilkina, Le Song. Learning
Combinatorial Optimization Algorithms over Graphs. NIPS 2017.
52. 52 / 54 KYOTO UNIVERSITY
参考える文献:
Wenqi Fan, Yao Ma, Qing Li, Yuan He, Eric Zhao, Jiliang Tang, Dawei Yin. Graph
Neural Networks for Social Recommendation. WWW 2019.
Maxime Gasse, Didier Chételat, Nicola Ferroni, Laurent Charlin, Andrea Lodi.
Exact Combinatorial Optimization with Graph Convolutional Neural Networks.
NeurIPS 2019.
Lauri Hella, Matti Järvisalo, Antti Kuusisto, Juhana Laurinharju, Tuomo
Lempiäinen, Kerkko Luosto, Jukka Suomela, Jonni Virtema. Weak Models of
Distributed Computing, with Connections to Modal Logic. PODC 2012.
Jiayi Huang, Mostofa Patwary, Gregory Diamos. Coloring Big Graphs with
AlphaGoZero. arXiv 2019.
53. 53 / 54 KYOTO UNIVERSITY
参考える文献:
Thomas N. Kipf, Max Welling. Semi-Supervised Classification with Graph
Convolutional Networks. ICLR 2017.
Haggai Maron, Ethan Fetaya, Nimrod Segol, Yaron Lipman. On the Universality
of Invariant Networks. ICML 2019.
Christopher Morris, Martin Ritzert, Matthias Fey, William L. Hamilton, Jan Eric
Lenssen, Gaurav Rattan, Martin Grohe. Weisfeiler and Leman Go Neural: Higher-
order Graph Neural Networks. AAAI 2019.
Namyong Park, Andrey Kan, Xin Luna Dong, Tong Zhao, Christos Faloutsos.
Estimating Node Importance in Knowledge Graphs Using Graph Neural
Networks. KDD 2019.
54. 54 / 54 KYOTO UNIVERSITY
参考える文献:
Ryoma Sato, Makoto Yamada, Hisashi Kashima. Approximation Ratios of Graph
Neural Networks for Combinatorial Problems. NeurIPS 2019.
Ryoma Sato, Makoto Yamada, Hisashi Kashima. Random Features Strengthen
Graph Neural Networks. arXiv 2020.
Nino Shervashidze, Pascal Schweitzer, Erik Jan van Leeuwen, Kurt Mehlhorn,
Karsten M. Borgwardt. Weisfeiler-Lehman Graph Kernels. JMLR 2011.
Petar Veličković, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Liò,
Yoshua Bengio. Graph Attention Networks. ICLR 2018.
Keyulu Xu, Weihua Hu, Jure Leskovec, Stefanie Jegelka. How Powerful are Graph
Neural Networks? ICLR 2019.