Seminar Schedule

Paper Pool

There are some papers collected HERE. You may want to select paper from them.

2017 Summer

We will have meeting on Friday 3:30PM at 542 or Lab.

  • Yu-Sheng
  • Wei-Lin
  • Li Chen
  • Roberts
  • Yu-Ting
  • Kent Loong
  • Pin-Yen
  • Chuan-Yao
  • Hung-yi
  • Jui-Nan

2017 Spring

We will have meeting on Friday 3:30PM at 542 or Lab.

  • 10/27 Jui-Nan The Marginal Value of Adaptive Gradient Methods in Machine Learning
  • 10/20 Wei-Lin and Yu-Sheng Their intern project in Alibaba
  • 10/13 Li Chen Naive Parallelization of Coordinate Descent Methods and an Application on Multi-core L1-regularized Classification
  • 09/29 Robert Understanding Machine Reading Comprehension
  • 09/22 Alex A Comparison of Optimization Methods for Large Scale Factorization Machine
  • 09/15 Yu-Ting On the Use of Stochastic Hessian Information in Optimization Methods for Machine Learning
  • 09/08 Kent Loong Identifying and attacking the saddle point problem in high-dimensional non-convex optimization
  • 09/01 Kent Loong Playing Atari with Deep Reinforcement Learning
  • 08/25 Pin-Yen Adam: A Method for Stochastic Optimization
  • 08/18 Chuan-Yao Large-scale Collaborative Ranking in Near-Linear Time
  • 08/11 Hung-yi Theory of the GMM Kernel
  • 08/04 Yu-Ting Trust Region Newton Method for Large-Scale Logistic Regression
  • 07/28 Chuan-Yao A Generic Coordinate Descent Framework for Learning from Implicit Feedback
  • 07/21 Hung-yi Spatial Decompositions for Large Scale SVMs
  • 07/14 Chi-Cheng His work.
  • 07/07 Kent Loong Scalable and Sustainable Deep Learning via Randomized Hashing
  • 06/30 Xiao-cong, Chih-Yang, sheng-Wei Their works.
  • 06/23 Alex Limited-memory Common-directions Method for Distributed Optimization and its Application on Empirical Risk Minimizatio
  • 06/16 Yu-Sheng liquidSVM: A Fast and Versatile SVM package
  • 06/09 Yu-Ting DiSMEC - Distributed Sparse Machines for Extreme Multi-label Classification
  • 06/02 Roberts On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima
  • 05/26 Chuan-Yao A Note on Decomposition Method for Support Vector Machine
  • 05/19 Chuan-Yao A Note on Decomposition Method for Support Vector Machine
  • 05/12 Wei-Lin (and Chih-Yang) would like to share their recent work about preconditioners for Newton-CG.7

2014 Fall

We will have meeting on Friday 3:30PM at 542 or Lab.

2013 Summer

We will have meeting on Tuesday and Thurday 2:30PM at 542 or Lab.

2013 Spring

We will have meeting on Friday 3PM at 542 or Lab.

  • 02/04 Ching-Pei presented his work on RankSVM.
  • 02/06 Ching-Pei finished presenting his RankSVM paper.
  • 03/01 Cheng-Hao will continue to present the above paper.
  • 03/08 Chun-Heng will continue to prove the remaining part of Trading Representability for Scalability: Adaptive Multi-Hyperplane Machine for Nonlinear Classification.
  • 03/015 Report the papers from the conference (1)
  • 03/22 Report the papers from the conference (2)
  • 03/29 Report the papers from the conference (3) and Po-Wei will present his comments from the reviewers.
  • 04/05 There will be no group meeting for the vacation.
  • 04/12 Report the papers from the conference (4)
  • 04/19 Tzu-Ming will present Cutting-Plane Training of Structural SVMs.
  • 04/26 Tzu-Ming will continue the presentation.
  • 05/03 Ching-pei will introduce the situation in Baidu and Tong will present his recent work.
  • 05/10 Tong will present his recent work.
  • 05/17 Tong will continue to present his recent work.
  • 05/24 Wei-Sheng will present Scalable Coordinate Descent Approaches to Parallel Matrix Factorization for Recommender Systems H.-F. Yu, C.-J. Hsieh et al., ICDM 2012 (Best paper)
  • 05/31 Yong will present the recent work by Yu-Chin and Wei-Sheng.
  • 06/07 Chien-Chih and Chun-Heng will present their recent work.
  • Wei-Lun
  • Wei-Cheng
  • Chengxia
  • Kai-Hsiang

2012 Fall

We will have meeting on Friday afternoon at 542 or at lab starting from 15:30

  • 9/21 Yu Tong will continue to present Towards Good Practice in Large-Scale Learning for Image Classification
  • 9/28 Po-Wei will present his work
  • 9/28 Wei-Cheng will present the reviewer's comments and the reply letter of Large-scale Linear Support Vector Regression
  • 10/5 Po-Wei will continue to present his work
  • 10/5 Wei-Cheng will continue to present the reviewer's comments and the reply letter of Large-scale Linear Support Vector Regression
  • 11/09 Cheng-xia Chang will present Recent Advances of Large-scale Linear Classification
  • 11/30 Tzu-Ming will present An Improved GLMNET for L1-regularized Logistic Regression and Support Vector Machines.
  • 12/7 Tzu-Ming will continue to present An Improved GLMNET for L1-regularized Logistic Regression and Support Vector Machines.
  • 12/14 Wei-Sheng will present Building support vector machines with reduced classifier complexity
  • 12/21 Wei-Sheng will continue to present Building support vector machines with reduced classifier complexity
  • 01/14 Wei-Cheng will present Large-scale Linear Support Vector Regression.
  • 01/16 Wei-Cheng will present Large-scale Linear Support Vector Regression.
  • 01/16 Po-Wei will present Complexity Analysis of Feasible Descent Methods for Convex Optimization
  • 01/18 Wei-Cheng will present Large-scale Linear Support Vector Regression.
  • 01/18 Po-Wei will present Complexity Analysis of Feasible Descent Methods for Convex Optimization
  • 01/21 Wei-Cheng will present Large-scale Linear Support Vector Regression.
  • 01/31 Chun-Heng will present Trading Representability for Scalability: Adaptive Multi-Hyperplane Machine for Nonlinear Classification.
  • Zhuang Yong
  • Ching-Pei

2012 Summer

  • 6/27 Tse-Ju presented Model Order Selection for Boolean Matrix Factorization.
  • 6/28 Tse-Ju continued to present Model Order Selection for Boolean Matrix Factorization.
  • 6/29 Tse-Ju finished presenting Model Order Selection for Boolean Matrix Factorization.
  • 7/4 Yu-Chin presented Title classification
  • 7/5 Yu-Chin finished Title classification
  • 7/5 Lei Weng-Chong presented Large-Scale Matrix Factorization with Distributed Stochastic Gradient Descent
  • 7/6 Lei Weng-Chong should presented Large-Scale Matrix Factorization with Distributed Stochastic Gradient Descent but canceled
  • 7/11 Lei Weng-Chong continued to present Large-Scale Matrix Factorization with Distributed Stochastic Gradient Descent
  • 7/12 Lei Weng-Chong should continued to present Large-Scale Matrix Factorization with Distributed Stochastic Gradient Descent but didn't show up
  • 7/13 Po-Wei presented and finished Selective Block Minimization for Faster Convergence of Limited Memory Large-scale Linear Models.
  • 7/18 Chieh-Yen presented Unbiased Offline Evaluation of Contextual-bandit-based News Article Recommendation Algorithms
  • 7/20 Ching-Pei finished Training linear ranking SVMs in linearithmic time using red-black trees.
  • 7/24 Cheng-Hao comes back to Taiwan.
  • 7/25 Lei Weng-Chong finished Large-Scale Matrix Factorization with Distributed Stochastic Gradient Descent
  • 7/25 Wei-Lun presented Low-degree Polynomial Mapping of NLP Data and Features Condensing by Hashing
  • 7/26 Wei-Lun finished Low-degree Polynomial Mapping of NLP Data and Features Condensing by Hashing
  • 7/27 Yu-Chin presented outlier detection techniques
  • 8/1 Ching-Pei presented How to Explain Individual Classification Decisions
  • 8/2 No meeting due to typhoon
  • 8/3 Ching-Pei finished How to Explain Individual Classification Decisions
  • 8/3 Chia-Hua presented Automatic Evaluation of Machine Translation Quality Using N-gram Co-Occurence Statistics.
  • 8/8 Chen-Hao presented Optimistic Bayesian Sampling in Contextual-Bandit Problems
  • 8/9 Chen-Haofinished Optimistic Bayesian Sampling in Contextual-Bandit Problems
  • 8/10 Po-Wei presented The Kernelized Stochastic Batch Perceptron
  • 8/15 Po-Wei finished The Kernelized Stochastic Batch Perceptron
  • 8/23 Wei-Lun continued to present Linear Support Vector Machines via Dual Cached Loops
  • 8/24 Yu-Chin presented A Dual Coordinate Descent Method for Large-scale Linear SVM
  • 8/29 Yu-Chin continued to present A Dual Coordinate Descent Method for Large-scale Linear SVM
  • 8/30 Yu-Chin continued to present A Dual Coordinate Descent Method for Large-scale Linear SVM
  • 8/31 Yu-Chin finished A Dual Coordinate Descent Method for Large-scale Linear SVM
  • 9/5 Zhuang Yong presented Trust region Newton method for large-scale logistic regression
  • 9/6 Zhuang Yong finished Trust region Newton method for large-scale logistic regression

2012 Spring

We will have meeting on Friday afternoon at 542 or at lab(we have a white board now!).

  • 2/24: KDD 2012 papers reviewing discussion
  • 3/2: KDD 2012 papers reviewing discussion
  • 3/9: KDD 2012 papers reviewing discussion
  • 3/16: ICML 2012 papers reviewing discussion
  • 3/23: ICML 2012 papeers reviewing discussion
  • 3/30: Yu-Chin presented Deep Learning via Hessian-free Optimization.
  • 4/6: Chia-Hua presented his SVR work.
  • 4/13: Chia-Hua continued to present his work.
  • 4/20: Chia-Hua finished presenting his work.
  • 4/27: Yu-Chin finished presenting Deep Learning via Hessian-free Optimization. Chia-Hua rehearsed his oral defense.
  • 5/4: Ching-pei presented his l2-loss multi-class SVM work. Chia-Hua will rehearse his oral defense again.
  • 5/11 Kai-Min presented his reinforcement learning work.
  • 5/18 Chien-Chih presented large-scale distributed machine learning.
  • 5/25 Chien-Chih continued to present large-scale distributed machine learning.

2012 Winter

We will have meeting on Monday, Thursday and Friday afternoon at 540/542/544/546.

2011 Fall

2011 Summer

2009 Fall

  • 9/25 (Fri.) Scan presetns Chunking with Support Vector Machines.
  • 10/2 (Fri.) Kai-Min presents Trust region Newton method for large-scale logistic regression
  • 10/9 (Fri.) Chen-Tse presents his work
  • 10/16 (Fri.) GuoXun present his work.
  • 10/23 (Fri.) GuoXun present his work.
  • 10/30 (Fri.) OMG present ???.
  • 11/6 (Fri.) PoPo present Fast Logistic Regression for Text Categorization with Variable-Length N-grams.
  • 11/20 (Fri.) Kai-Min present An Affine-Scaling Interior-Point Method For Continuous Knapsack
  • 11/27 (Fri.) Kai-Min present An Affine-Scaling Interior-Point Method For Continuous Knapsack
  • 12/4 (Fri.) Scan present An Agence web paris Affine-Scaling Interior-Point Method For Continuous Knapsack Constraints.

2009 Spring

We will have meeting on Friday afternoon at 524/526, starting from 3:20pm.

  • 5/1 (Fri.) Scan presents the rest of SMO linear convergence.
  • 5/8 (Fri.) Kai-Min presents Sparse Online Learning via Truncated Gradient.
  • 5/14 (Thu.) Janet presetns her work “Maximum Entropy”.
  • 5/21 (Thu.) Ho Chia-Hua presents A Dual Coordinate Descent Method for Large-scale Linear SVM
  • 6/5 (Fri.) Kai-Min presents Yin-Wen's thesis.
  • 5/29 (Thu.) Janet presents pipibjc's thesis.
  • 6/4 (Thu.) Kai-Min presents Yin-Wen's thesis.
  • 6/5 (Fri.) Chen-Tse,博瀚 presents eaudex's thesis.
  • 6/12 (Fri.) Kai-Min presetns the rest of Sparse Online Learning via Truncated Gradient.
  • 6/19 (Fri.) Tian-Liang presetns Learning to Classify with Missing and Corrupted Features .
  • 6/26 (Fri.) pipibjc presetns LYin-Wen's thesis.
  • 7/2 (Thu.) Tian-Liang presetns the rest of Learning to Classify with Missing and Corrupted Features .
  • 7/9 (Thu.) Dino presents his research
  • 7/10 (Fri.) Chia Hua presents An Interior-Point Method for Large-Scale l1-Regularized Logistic Regression
  • 7/16 (Thu.) Scan presetns Stochastic Methods for L1 Regularized Loss Minimization.
  • 7/17 (Fri.) Chia Hua presents An Interior-Point Method for Large-Scale l1-Regularized Logistic Regression
  • 7/23 (Thu.) 2nd-year master students presents their work
  • 7/24 (Fri.) Others present their work
  • 7/30 (Thu.) Chen-Tse 博瀚 presents svvd
  • 8/6 (Thu.) Kai-Min presents A Note on Decomposition Methods for Support Vector Regression
  • 8/7 (Fri.) Typhoon 莫拉克
  • 8/13 (Thu.) Chia Hua presents arge Linear Classification when Data Cannot Fit in Memory
  • 8/14 (Fri.) Tian-Liang presents Learning Linear Dynamical Systems without Sequence Information
  • 8/21 (Fri.) Chen-Tse presents Tree Decomposition for Large-Scale Support Vector Machines
  • 8/27 (Thu.) Scan presents Hash Kernels for Structured Data
  • 8/28 (Fri.) Po-Han presents Feature Hashing for Large Scale Multitask Learning
  • 9/3 (Thu.) Kai-Min presents Trust region Newton method for large-scale logistic regression
  • 9/4 (Fri.) Tian-Liang presents
  • 9/10 (Thu.) Chen-Tse presents Dual Augmented Lagrangian Method for Efficent Sparse Reconstruction
  • 9/11 (Fri.) Po-Han presents Fast Solution of L1-morm Minimization Problems When the Solution May be Sparse
  • ?? (??) ?? Bundle Methods for Regularized Risk Minimization.

2007 Fall

We will have meeting on Monday afternoon at 524/526, starting from 3:00pm.

  • 11/05(Mon.) Yin-Wen presents about structure svm.
  • 10/29(Mon.) Yin-Wen presents about structure svm.
  • 10/22(Mon.) No meeting today because of a speech.
  • 10/15(Mon.) pipibjc presents about structure svm on LETOR.
  • 10/08(Mon.) biconnect, rainfarmer present about their research.

2007 Summer

We will have meeting on Friday morning at 524/526, starting from 10:00 am.

  • 9/14(Fri.): GuoXun present about pegasos.
  • 9/07(Fri.): GuoXun present “A Study on Threshold Selection for Multilable Classification”.
  • 8/31(Fri.): everyone talk about their current research work.
  • 8/24(Fri.): biconnect present about gaussian process regression.
  • 8/24(Fri.): Bee present Chapter 3.
  • 8/17(Fri.): canceled because of typhoon.
  • 8/10(Fri.): Ma present about Image thumbnailing.
  • 8/03(Fri.): Bee present Chapter 3.
  • 8/03(Fri.): Rofu present paper.
  • 7/27(Fri.): GuoXun present the rest part of Chapter 2.
  • 7/27(Fri.): pipibjc present about LETOR.
  • 7/20(Fri.): GuoXun present Chapter 2 of Numerical Optimization, from page 11 to page 22.
  • 7/20(Fri.): greenoyster present about his research.
  • 7/13(Fri.): Ma presents netflix challenge.

2007 Spring

We will have meeting on Friday morning at 524, starting from 9:30.

  • 6/8(Fri.): biconnect/rainfermer present hotterornotter
  • 6/1(Fri.): rafan present his Master thesis
  • 5/18(Fri.): GuoXun present sub-GrAdient SOlver for SVM
  • 5/18(Fri.): acherub present comments on CVM
  • 5/11(Fri.): acherub present CVM
  • 5/4(Fri.): zao present Trust Region Newton Method for Large-Scale Regularized Logistic Regression
  • 4/20(Fri.): oyster present Trust Region Newton Method for Large-Scale Regularized Logistic Regression
  • 3/30(Fri.): pipibjc presents the python interface of libsvm
  • 3/23(Fri.): Ma presents Statistical Edge Detection: Learning and Evaluating Edge Cues
  • 3/16(Fri.): Ma presents Statistical Edge Detection: Learning and Evaluating Edge Cues
  • 3/16(Fri.): GuoXun presents libsvm string code

2007 Winter

We will have meeting on Tuesday and Friday starting from 13:00.

2006 Fall

We will have meeting on Friday afternoon at 524 this semester, starting from 12:30 PM.

2006 Summer

We will have meeting on Friday morning at 524 this summer, starting from 9:30 AM.

2006 Spring

We will have meeting on Friday morning at 524 this semester.

2006 Winter

We will have meeting at 524 on Monday and Friday afternoon during the vacation.

  • 2/18(Fri.): Tzu-Kuo presents
  • 2/13(Mon.): puffer presents
  • 2/10(Fri.): puffer presents
  • 2/06(Mon.): Tzu-Kuo presents

2005 Fall

We will have meetings at 524 on Tuesday morning/Friday afternoon this semester.

 
seminar.txt · Last modified: 2017/10/26 12:43 by r05922038     Back to top