November 2017 - Dm-meeting - mailman.cs.vt.edu

Thursday, Nov 30th, @ 11:00 am, Torg 3160 A, Xinfeng Xu
by Sorour Ekhtiari Amiri 30 Nov '17

30 Nov '17

Hi everyone, Please join us for our next DM-Meeting. ****DM-Meeting Fall 2017**** WHO: Xinfeng Xu WHEN: Thursday, Nov. 30th, 2017, @ 11:00 am WHERE: Torg 3160 A WHAT: Xinfeng will present the following papers, 1. EmbedJoin: Efficient Edit Similarity Joins via Embeddings: Haoyu Zhang, Qin Zhang, KDD17 Abstract: We study the problem of edit similarity joins, where given a set of strings and a threshold value K, we want to output all pairs of strings whose edit distances are at most K. Edit similarity join is a fundamental problem in data cleaning/integration, bioinformatics, collaborative filtering and natural language processing, and has been identified as a primitive operator for database systems. This problem has been studied extensively in the literature. However, we have observed that all the existing algorithms fall short on long strings and large distance thresholds. In this paper we propose an algorithm named EmbedJoin which scales very well with string length and distance threshold. Our algorithm is built on the recent advance of metric embeddings for edit distance, and is very different from all of the previous approaches. We demonstrate via an extensive set of experiments that EmbedJoin significantly outperforms the previous best algorithms on long strings and large distance thresholds. 2. Efficient Correlated Topic Modeling with Topic Embedding, Junxian He, Zhiting Hu, Taylor Berg-Kirkpatrick, Ying Huang, Eric P. Xing, KDD17 Abstract: Correlated topic modeling has been limited to small model and problem sizes due to their high computational cost and poor scaling. In this paper, we propose a new model which learns compact topic embeddings and captures topic correlations through the closeness between the topic vectors. Our method enables efficient inference in the low-dimensional embedding space, reducing previous cubic or quadratic time complexity to linear w.r.t the topic size. We further speedup variational inference with a fast sampler to exploit sparsity of topic occurrence. Extensive experiments show that our approach is capable of handling model and data scales which are several orders of magnitude larger than existing correlation results, without sacrificing modeling quality by providing competitive or superior performance in document classification and retrieval. Best regards, Sorour

2 1

No Meeting this week
by B. Aditya Prakash 14 Nov '17

14 Nov '17

Hi everyone, I will be out, so there won't be any DM-meeting this week. Nevertheless, if you guys want to meet informally at the usual time, please feel free to do so, and talk about travels, papers and other things. We can order pizza as well. cheers, - Aditya

1 0

Congratulations to Yao---ICDM Best-paper candidate
by B. Aditya Prakash 14 Nov '17

14 Nov '17

Congratulations to Yao, as his paper 'Data-driven Immunization' was a 'Best-paper candidate' in the upcoming ICDM 2017 conference! cheers, - Aditya

1 0

Thursday, Nov 9th, @ 11:00 am, Torg 3160 A, Bijaya Adhikari and Sorour Amiri
by Sorour Ekhtiari Amiri 06 Nov '17

06 Nov '17

Hi everyone, Please join us for our next DM-Meeting. ****DM-Meeting Fall 2017**** WHO: Bijaya Adhikari and Sorour Amiri WHEN: Thursday, Nov. 9th, 2017, @ 11:00 am WHERE: Torg 3160 A WHAT: Bijaya and I will give a brief tutorial about Representation Learning on Graphs and deep reinforcement learning. Title: Deep reinforcement learning Abstract: A major goal of artificial intelligence is to create general-purpose agents that can perform effectively in a wide range of challenging tasks. To achieve this goal, it is necessary to combine reinforcement learning (RL) agents with powerful and flexible representations. The key idea of deep RL is to use neural networks to provide this representational power. In this tutorial, we will present a family of algorithms in which deep neural networks are used for value functions, policies, or environment models. State-of-the-art results will be presented in a variety of domains, including Atari games, 3D navigation tasks, continuous control domains and the game of Go. Title: Representation Learning on Graphs: Methods and Applications Abstract: Machine learning on graphs is an important and ubiquitous task with applications ranging from drug design to friendship recommendation in social networks. The primary challenge in this domain is finding a way to represent, or encode, graph structure so that it can be easily exploited by machine learning models. Traditionally, machine learning approaches relied on user-defined heuristics to extract features encoding structural information about a graph (e.g., degree statistics or kernel functions). However, recent years have seen a surge in approaches that automatically learn to encode graph structure into low-dimensional embeddings, using techniques based on deep learning and nonlinear dimensionality reduction. Here we provide a conceptual review of key advancements in this area of representation learning on graphs, including matrix factorization-based methods, random-walk based algorithms, and graph convolutional networks. We review methods to embed individual nodes as well as approaches to embed entire (sub)graphs. In doing so, we develop a unified framework to describe these recent approaches, and we highlight a number of important applications and directions for future work. Best regards, Sorour

1 0