Knowledge Transfer with Interactive Learning of Semantic Relationships
|
|
|
- Eugenia Nash
- 5 years ago
- Views:
Transcription
1 Knowledge Transfer with Interactive Learning of Semantic Relationships Feb. 15, 2016 Jonghyun Choi Sung Ju Hwang, Leonid Sigal and Larry S. Davis University of Maryland Institute of Advanced Computer Studies (UMIACS) Comcast Labs, DC Ulsan National Institute of Science and Technology (UNIST) Disney Research, Pittsburgh
2 We Can Recognize This Image courtesy by
3 We Can Recognize This Too Image courtesy by
4 And We Can Easily Infer What This Is [1] J. Feldman, The structure of perceptual categories, Journal of Mathematical Psychology 1997 [2] J. Tenenbaum, Bayesian modeling of human concept learning, NIPS 1999 [3] T. Tommasi et al, Safety in Numbers: Learning Categories from Few Examples with Multi Model Knowledge Transfer, CVPR 2010 [4] Qi et al, Towards Cross-Category Knowledge Propagation for Learning Visual Concepts, CVPR 2011 [5] B. Lake, R. Salakhutdinov, J. Tenenbaum, Human-level concept learning through probabilistic program induction, Science 2015 Image courtesy by
5 Knowing Some Categories Allows Recognition of Others If we know these categories well We may recognize a bird-dog better [1] Image courtesy by Google [1] Tommasi, T.; Orabona, F.; and Caputo, B.. Safety in Numbers: Learning Categories from Few Examples with Multi Model Knowledge Transfer. CVPR 2010
6 Knowing Some Categories Allows Recognition of Others If we know these categories well We may recognize a bird-dog better [1] Relationship Image courtesy by Google [1] Tommasi, T.; Orabona, F.; and Caputo, B.. Safety in Numbers: Learning Categories from Few Examples with Multi Model Knowledge Transfer. CVPR 2010
7 Knowing Some Categories Allows Recognition of Others Anchor category If we know these categories well We may recognize a bird-dog better [1] Relationship Image courtesy by Google [1] Tommasi, T.; Orabona, F.; and Caputo, B.. Safety in Numbers: Learning Categories from Few Examples with Multi Model Knowledge Transfer. CVPR 2010
8 Knowing Some Categories Allows Recognition of Others Anchor category If we know these categories well Target categories We may recognize a bird-dog better [1] Relationship Image courtesy by Google [1] Tommasi, T.; Orabona, F.; and Caputo, B.. Safety in Numbers: Learning Categories from Few Examples with Multi Model Knowledge Transfer. CVPR 2010
9 Learning Semantic Relationship in Metric To improve the classification accuracy of rarely seen object category in metric learning framework
10 Learning Semantic Relationship in Metric To improve the classification accuracy of rarely seen object category in metric learning framework Target category
11 What Relationship? A is more similar to B than C Chimpanzee is more similar to Gorilla than Deer
12 What Relationship? Target category Anchor categories A is more similar to B than C Chimpanzee is more similar to Gorilla than Deer
13 What Relationship? Target category Anchor categories A is more similar to B than C Chimpanzee is more similar to Gorilla than Deer How to obtain the relationships?
14 Creating Relational Knowledge Base is expensive - Exponentially many (N 3 ) knowledges - Difficult to answer Ambiguous relative similarity - Every relationship is not equally useful Some are more useful, some are less or useless
15 Creating Relational Knowledge Base is expensive - Exponentially many (N 3 ) knowledges - Difficult to answer Ambiguous relative similarity - Every relationship is not equally useful Some are more useful, some are less or useless
16 Creating Relational Knowledge Base is expensive - Exponentially many (N 3 ) knowledges - Difficult to answer Ambiguous relative similarity - Every relationship is not equally useful Some are more useful, some are less or useless
17 Creating Relational Knowledge Base is expensive - Exponentially many (N 3 ) knowledges - Difficult to answer Ambiguous relative similarity - Every relationship is not equally useful Some are more useful, some are less or useless
18 Ambiguous Relationship Who is more similar to Ironman? Image courtesy by Disney and Toei Company 9
19 Ambiguous Relationship Who is more similar to Ironman? Power ranger? Image courtesy by Disney and Toei Company 9
20 Ambiguous Relationship Who is more similar to Ironman? Power ranger? Image courtesy by Disney and Toei Company 9 Tony Stark?
21 Not Every Relationship is Equally Useful for Classification Classifier for prettiness Snow-white Image courtesy by Disney
22 Not Every Relationship is Equally Useful for Classification Classifier for prettiness Snow-white > Queen Dwarf >> Snow-white is prettier than the Queen rather than the Seven dwarfs? Image courtesy by Disney
23 Not Every Relationship is Equally Useful for Classification Classifier for prettiness Rapunzel < Snow-white > Queen << Cinderella >> Dwarf Snow-white is prettier than the Queen rather than the Seven dwarfs? Snow-white is prettier than Rapunzel rather than Cinderella? Image courtesy by Disney
24 Our Approach: Ask A Few Useful Questions
25 Our Approach: Ask A Few Useful Questions Ask most useful questions in interactions
26 We Propose To learn a semantic space for target categories with constraints of semantic distance from the anchor categories Target category: one to improve the classification with few samples Anchor category: one to transfer knowledge from with more samples
27 We Propose To learn a semantic space for target categories with constraints of semantic distance from the anchor categories Example: - Chimpanzee is more similar to Gorilla than Deer? - Collie is more similar to Dalmatian than Giraffe?
28 Learning Classification Embeddings
29 Learning Classification Embeddings Training images of Anchor Categories
30 Learning Classification Embeddings Training images of Anchor Categories Training images of Target Categories
31 Learning Classification Embeddings Training images of Anchor Categories Training images of Target Categories W W
32 Learning Classification Embeddings Training images of Anchor Categories Training images of Target Categories W W Gorilla Deer Anchor Categories Giraffe Dalmatian
33 Learning Classification Embeddings Training images of Anchor Categories Training images of Target Categories W W Gorilla Deer Chimpanzee Anchor Categories Giraffe Dalmatian Collie Target Categories
34 Learning Classification Embeddings Training images of Anchor Categories Training images of Target Categories W W Gorilla u a1 Deer u a2 Chimpanzee u t1 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories
35 Refine Classification Embedding by Relational Semantics W W Gorilla u a1 Deer u a2 Chimpanzee u t1 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories
36 Refine Classification Embedding by Relational Semantics W W Gorilla u a1 Deer u a2 Chimpanzee u t1 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
37 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Anchor Categories Giraffe u a3 Deer u a2 Dalmatian u a4 Chimpanzee u t1 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
38 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Anchor Categories Giraffe u a3 Deer u a2 Dalmatian u a4 Chimpanzee u t1 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
39 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Deer u a2 Chimpanzee u t1 u t1 u a1 < u t1 u a2 u t2 u a3 < u t2 u a4 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
40 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Deer u a2 Chimpanzee u t1 u t1 u a1 < u t1 u a2 u t2 u a3 < u t2 u a4 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
41 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Deer u a2 Chimpanzee u t1 u t1 u a1 < u t1 u a2 u t2 u a3 < u t2 u a4 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
42 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Deer u a2 Chimpanzee u t1 u t1 u a1 < u t1 u a2 u t2 u a3 < u t2 u a4 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
43 Learning Classification Embedding by Relational Semantics Feedback W W Gorilla u a1 Deer u a2 Chimpanzee u t1 u t1 u a1 < u t1 u a2 u t2 u a3 < u t2 u a4 Anchor Categories Giraffe u a3 Dalmatian u a4 Collie u t2 Target Categories? - Is Chimpanzee more similar to Gorilla than Deer? - Is Collie more similar to Dalmatian than Giraffe?
44 Relationship Transferred Metric Learning (RTML) Embed both visual feature and label entities - First on the Anchor categories only min W A,U A N A X i=1 X c2c A L W A, x i, u c + 1 kw A k 2 F + 2 ku A k 2 F, s.t. L(W A, x i, u c )= max kw A x i u yi k 2 2 kw A x i u c k , 0, 8i, 8c 6= y i W: subspace mapper, U: label embeddings, [1] K. Weinberger and L. Saul, Distance Metric Learning for Large Margin Nearest Neighbor Classification, JMLR 2009
45 Relationship Transferred Metric Learning (RTML) Embed both visual feature and label entities - First on the Anchor categories only min W A,U A Large margin embedding [1] on Anchor Categories N A X i=1 X c2c A L W A, x i, u c + 1 kw A k 2 F + 2 ku A k 2 F, s.t. L(W A, x i, u c )= max kw A x i u yi k 2 2 kw A x i u c k , 0, 8i, 8c 6= y i W: subspace mapper, U: label embeddings, [1] K. Weinberger and L. Saul, Distance Metric Learning for Large Margin Nearest Neighbor Classification, JMLR 2009
46 Relationship Transferred Metric Learning (RTML) Learn target category label prototypes and update W min W T,U T N T X i=1 X c2c T L W T, x i, u c + 3 kw T W A k 2 F + X j + 1 kw T k 2 F + 2 kuk 2 F (R j, U), s.t. L(W T, x i, u c ) = max kw T x i u yi k 2 2 kw T x i u c k , 0, 8i, 8c 6= y i, R j R W: subspace mapper, U: label embeddings, : semantic relational constraints [1] K. Weinberger and L. Saul, Distance Metric Learning for Large Margin Nearest Neighbor Classification, JMLR 2009
47 Relationship Transferred Metric Learning (RTML) Learn target category label prototypes and update W min W T,U T Large margin embedding [1] on Target Categories N T X i=1 X c2c T L W T, x i, u c + 3 kw T W A k 2 F + X j + 1 kw T k 2 F + 2 kuk 2 F (R j, U), s.t. L(W T, x i, u c ) = max kw T x i u yi k 2 2 kw T x i u c k , 0, 8i, 8c 6= y i, R j R W: subspace mapper, U: label embeddings, : semantic relational constraints [1] K. Weinberger and L. Saul, Distance Metric Learning for Large Margin Nearest Neighbor Classification, JMLR 2009
48 Relationship Transferred Metric Learning (RTML) Learn target category label prototypes and update W min W T,U T Large margin embedding [1] on Target Categories N T X i=1 X c2c T L W T, x i, u c + 1 kw T k 2 F + 2 kuk 2 F + 3 kw T W A k 2 F + X (R j, U), j s.t. L(W T, x i, u c ) = max kw T x i u yi k 2 2 kw T x i u c k , 0, 8i, 8c 6= y i, R j R Enforce semantic relation W: subspace mapper, U: label embeddings, : semantic relational constraints [1] K. Weinberger and L. Saul, Distance Metric Learning for Large Margin Nearest Neighbor Classification, JMLR 2009
49 Relationship Transferred Metric Learning (RTML) Learn target category label prototypes and update W min W T,U T Large margin embedding [1] on Target Categories N T X i=1 X c2c T L W T, x i, u c + 1 kw T k 2 F + 2 kuk 2 F + 3 kw T W A k 2 F + X (R j, U), j Bound to W A s.t. L(W T, x i, u c ) = max kw T x i u yi k 2 2 kw T x i u c k , 0, 8i, 8c 6= y i, R j R Enforce semantic relation W: subspace mapper, U: label embeddings, : semantic relational constraints [1] K. Weinberger and L. Saul, Distance Metric Learning for Large Margin Nearest Neighbor Classification, JMLR 2009
50 Semantic Relational Constraints u t Chimpanzee u a2 Deer u a1 Gorilla ku a1 u t k 2 2 < ku a2 u t k < ku a u 2 t k 2 2 ku u a1 t k 2 2 ut should be closer to ua1 than ua2! min U 0 < 1 max 1 ku a2 u t k 2 2 ku a1 u t k 2 2 ku a2 u t k 2 2 ku a1 u t k 2 2, 0.
51 Semantic Relational Constraints min U max 1 ku a2 u t k 2 2 ku a1 u t k 2 2, 0. ut should be closer to ua1 than ua2 Neither convex nor differentiable - So, relax by a way of [1]: X j (R j,u) = 1 h ku a1 u t k 2 2 ku a2 u t k 2 2 [1] Hwang et al., Analogy-preserving semantic embed- ding for visual object categorization, ICML 2013
52 Semantic Relational Constraints min U max 1 ku a2 u t k 2 2 ku a1 u t k 2 2, 0. ut should be closer to ua1 than ua2 Neither convex nor differentiable - So, relax by a way of [1]: X j (R j,u) = 1 h ku a1 u t k 2 2 ku a2 u t k 2 2 Objective function now becomes convex [1] Hwang et al., Analogy-preserving semantic embed- ding for visual object categorization, ICML 2013
53 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement Anchors How to obtain the scores 1. No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
54 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement Anchors How to obtain the scores 1. No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
55 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement Anchors How to obtain the scores 1. No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
56 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement Anchors How to obtain the scores 1. No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
57 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement How to obtain the scores Anchors No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
58 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement How to obtain the scores Anchors No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
59 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement Anchors How to obtain the scores 1. No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
60 Questions in What Order? Ones improving the accuracy the most - Score the questions by expected accuracy improvement Anchors How to obtain the scores 1. No scores - Random (baseline) 2. Training accuracy (entropy) Targets 3. Accuracy improvement on a validation set 4. Linear regression: predict the accuracy improvement on the validation set
61 Experiments Datasets Animals with Attributes (AwA) [1] - 50 animal classes (30,475 images) - 10 target classes (2,5,10 training/class) - 40 anchor classes (30 training/class) ImageNet-50 [2] - 50 object classes (70,380 images) - 10 target classes (2,5,10 training/class) - 40 anchor classes (30 training/class) [1] C. H. Lampert, H. Nickisch, and S. Harmeling. "Learning To Detect Unseen Object Classes by Between-Class Attribute Transfer". CVPR, 2009 [2] S. J. Hwang, Analogy-preserving Semantic Embedding for Visual Object Categorization, ICML 2013
62 Questions Generated At Iterations Top ranked query for persian-cat at each iteration # Iteration Positively answered query at its highest rank 1 fox - persian cat < blue whale - persian cat 2 grizzly bear - persian cat < horse - persian cat 3 dalmatian - persian cat < beaver - persian cat 4 dalmatian - persian cat < german shepherd - persian cat
63 Questions Generated At Iterations Top ranked query for persian-cat at each iteration # Iteration Positively answered query at its highest rank 1 fox - persian cat < blue whale - persian cat 2 grizzly bear - persian cat < horse - persian cat 3 dalmatian - persian cat < beaver - persian cat 4 dalmatian - persian cat < german shepherd - persian cat
64 Questions Generated At Iterations Top ranked query for persian-cat at each iteration # Iteration Positively answered query at its highest rank 1 fox - persian cat < blue whale - persian cat 2 grizzly bear - persian cat < horse - persian cat 3 dalmatian - persian cat < beaver - persian cat 4 dalmatian - persian cat < german shepherd - persian cat
65 Questions Generated At Iterations Top ranked query for persian-cat at each iteration # Iteration Positively answered query at its highest rank 1 fox - persian cat < blue whale - persian cat 2 grizzly bear - persian cat < horse - persian cat 3 dalmatian - persian cat < beaver - persian cat 4 dalmatian - persian cat < german shepherd - persian cat
66 Questions Generated At Iterations Top ranked query for persian-cat at each iteration # Iteration Positively answered query at its highest rank 1 fox - persian cat < blue whale - persian cat 2 grizzly bear - persian cat < horse - persian cat 3 dalmatian - persian cat < beaver - persian cat 4 dalmatian - persian cat < german shepherd - persian cat As interactions continue, top ranked query becomes semantically more meaningful.
67 Qualitative Results Persian'Cat' Nearest' Nearest'Neighbors'in'Anchor'Classes' RTML Ours'(iter'3)' Wolf' Siamese3Cat' Fox' Weasel' Hamster' RTML RTML Ours'(iter'2)' Wolf' Siamese3Cat' Hamster' Chihuahua' Weasel' RTML Ours'(iter'1)' Wolf' Hamster' Siamese3Cat' Fox' Weasel' Blue3Whale' Dolphin' Hamster' Siamese3Cat' Wolf' LME3Transfer' RTML As iterations Blue3Whale' proceed, Dolphin' Hamster' Siamese3Cat' nearest Wolf' neighbors LME' become more semantically meaningful
68 Qualitative Results Persian'Cat' Nearest' Nearest'Neighbors'in'Anchor'Classes' RTML Ours'(iter'3)' Wolf' Siamese3Cat' Fox' Weasel' Hamster' RTML RTML Ours'(iter'2)' Wolf' Siamese3Cat' Hamster' Chihuahua' Weasel' RTML Ours'(iter'1)' Wolf' Hamster' Siamese3Cat' Fox' Weasel' Blue3Whale' Dolphin' Hamster' Siamese3Cat' Wolf' LME3Transfer' RTML As iterations Blue3Whale' proceed, Dolphin' Hamster' Siamese3Cat' nearest Wolf' neighbors LME' become more semantically meaningful
69 Qualitative Results Persian'Cat' Nearest' Nearest'Neighbors'in'Anchor'Classes' RTML Ours'(iter'3)' Wolf' Siamese3Cat' Fox' Weasel' Hamster' RTML RTML Ours'(iter'2)' Wolf' Siamese3Cat' Hamster' Chihuahua' Weasel' RTML Ours'(iter'1)' Wolf' Hamster' Siamese3Cat' Fox' Weasel' Blue3Whale' Dolphin' Hamster' Siamese3Cat' Wolf' LME3Transfer' RTML As iterations Blue3Whale' proceed, Dolphin' Hamster' Siamese3Cat' nearest Wolf' neighbors LME' become more semantically meaningful
70 Qualitative Results Persian'Cat' Nearest' Nearest'Neighbors'in'Anchor'Classes' RTML Ours'(iter'3)' Wolf' Siamese3Cat' Fox' Weasel' Hamster' RTML RTML Ours'(iter'2)' Wolf' Siamese3Cat' Hamster' Chihuahua' Weasel' RTML Ours'(iter'1)' Wolf' Hamster' Siamese3Cat' Fox' Weasel' Blue3Whale' Dolphin' Hamster' Siamese3Cat' Wolf' LME3Transfer' RTML As iterations Blue3Whale' proceed, Dolphin' Hamster' Siamese3Cat' nearest Wolf' neighbors LME' become more semantically meaningful
71 Qualitative Results Persian'Cat' Nearest' Nearest'Neighbors'in'Anchor'Classes' RTML Ours'(iter'3)' Wolf' Siamese3Cat' Fox' Weasel' Hamster' RTML RTML Ours'(iter'2)' Wolf' Siamese3Cat' Hamster' Chihuahua' Weasel' RTML Ours'(iter'1)' Wolf' Hamster' Siamese3Cat' Fox' Weasel' Blue3Whale' Dolphin' Hamster' Siamese3Cat' Wolf' LME3Transfer' RTML As iterations Blue3Whale' proceed, Dolphin' Hamster' Siamese3Cat' nearest Wolf' neighbors LME' become more semantically meaningful
72 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy faster
73 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy faster batchmode question generation
74 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy question faster batchmode question generation Interactive generation
75 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy question faster batchmode question generation Interactive generation
76 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy question faster batchmode question generation Interactive generation
77 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy question faster batchmode question generation Interactive generation When to stop?
78 Reduced Number of Questions Answer to better questions at every iteration - improve accuracy question faster batchmode question generation Interactive generation When to stop? : start to decrease in two consecutive iterations
79 By Different Query Selection Criteria #samples/class Animals with Attribute LME 22.51± ± ±1.33 LME-Transfer 24.59± ± ±1.67 Random 24.75± ± ±1.66 Entropy 24.96± ± ±1.91 Active-Regression 25.43± ± ±0.88 Active 26.62± ± ±1.33 Interactive 27.24± ± ±1.60 Interactive-UB 28.57± ± ±1.83 ImangeNet-50 LME 23.20± ± ±1.62 LME-Transfer 23.47± ± ±1.03 Random 24.23± ± ±2.26 Entropy 24.60± ± ±0.99 Active-Regression 23.34± ± ±0.89 Active 24.35± ± ±1.01 Interactive 24.95± ± ±1.01 Interactive-UB 25.15± ± ±1.53 Classification Accuracy (%) for Comparing Quality of Scoring Function
80 By Different Query Selection Criteria #samples/class Animals with Attribute LME 22.51± ± ±1.33 LME-Transfer 24.59± ± ±1.67 Random 24.75± ± ±1.66 Entropy 24.96± ± ±1.91 Active-Regression 25.43± ± ±0.88 Active 26.62± ± ±1.33 Interactive 27.24± ± ±1.60 Interactive-UB 28.57± ± ±1.83 ImangeNet-50 LME 23.20± ± ±1.62 LME-Transfer 23.47± ± ±1.03 Random 24.23± ± ±2.26 Entropy 24.60± ± ±0.99 Active-Regression 23.34± ± ±0.89 Active 24.35± ± ±1.01 Interactive 24.95± ± ±1.01 Interactive-UB 25.15± ± ±1.53 Baseline Proposed Classification Accuracy (%) for Comparing Quality of Scoring Function
81 Summary Propose an efficient and interactive strategy for collecting category relationship semantics Embedding semantic regularization into a classification model ensures the model to be semantically more meaningful over iterations Improve classification accuracy with small number of human verifications
82 Q/A Thank you!
83 Summary Propose an efficient and interactive strategy for collecting category relationship semantics Embedding semantic regularization into a classification model ensures the model to be semantically more meaningful over iterations Improve classification accuracy with small number of human verifications
Fantope Regularization in Metric Learning
Fantope Regularization in Metric Learning CVPR 2014 Marc T. Law (LIP6, UPMC), Nicolas Thome (LIP6 - UPMC Sorbonne Universités), Matthieu Cord (LIP6 - UPMC Sorbonne Universités), Paris, France Introduction
Using Both Latent and Supervised Shared Topics for Multitask Learning
Using Both Latent and Supervised Shared Topics for Multitask Learning Ayan Acharya, Aditya Rawal, Raymond J. Mooney, Eduardo R. Hruschka UT Austin, Dept. of ECE September 21, 2013 Problem Definition An
Semi Supervised Distance Metric Learning
Semi Supervised Distance Metric Learning wliu@ee.columbia.edu Outline Background Related Work Learning Framework Collaborative Image Retrieval Future Research Background Euclidean distance d( x, x ) =
Style-aware Mid-level Representation for Discovering Visual Connections in Space and Time
Style-aware Mid-level Representation for Discovering Visual Connections in Space and Time Experiment presentation for CS3710:Visual Recognition Presenter: Zitao Liu University of Pittsburgh ztliu@cs.pitt.edu
Stable and Efficient Representation Learning with Nonnegativity Constraints. Tsung-Han Lin and H.T. Kung
Stable and Efficient Representation Learning with Nonnegativity Constraints Tsung-Han Lin and H.T. Kung Unsupervised Representation Learning Layer 3 Representation Encoding Sparse encoder Layer 2 Representation
A Randomized Approach for Crowdsourcing in the Presence of Multiple Views
A Randomized Approach for Crowdsourcing in the Presence of Multiple Views Presenter: Yao Zhou joint work with: Jingrui He - 1 - Roadmap Motivation Proposed framework: M2VW Experimental results Conclusion
Generalisation and inductive reasoning. Computational Cognitive Science 2014 Dan Navarro
Generalisation and inductive reasoning Computational Cognitive Science 2014 Dan Navarro Where are we at? Bayesian statistics as a general purpose tool for doing inductive inference A specific Bayesian
Large-scale Image Annotation by Efficient and Robust Kernel Metric Learning
Large-scale Image Annotation by Efficient and Robust Kernel Metric Learning Supplementary Material Zheyun Feng Rong Jin Anil Jain Department of Computer Science and Engineering, Michigan State University,
Linear Classifiers IV
Universität Potsdam Institut für Informatik Lehrstuhl Linear Classifiers IV Blaine Nelson, Tobias Scheffer Contents Classification Problem Bayesian Classifier Decision Linear Classifiers, MAP Models Logistic
Metric Learning. 16 th Feb 2017 Rahul Dey Anurag Chowdhury
Metric Learning 16 th Feb 2017 Rahul Dey Anurag Chowdhury 1 Presentation based on Bellet, Aurélien, Amaury Habrard, and Marc Sebban. "A survey on metric learning for feature vectors and structured data."
Classification Part 2. Overview
Classification Part 2 Dr. Sanjay Ranka Professor Computer and Information Science and Engineering University of Florida, Gainesville Overview Rule based Classifiers Nearest-neighbor Classifiers Data Mining
Machine Learning Basics Lecture 7: Multiclass Classification. Princeton University COS 495 Instructor: Yingyu Liang
Machine Learning Basics Lecture 7: Multiclass Classification Princeton University COS 495 Instructor: Yingyu Liang Example: image classification indoor Indoor outdoor Example: image classification (multiclass)
Neural Networks. Single-layer neural network. CSE 446: Machine Learning Emily Fox University of Washington March 10, /9/17
3/9/7 Neural Networks Emily Fox University of Washington March 0, 207 Slides adapted from Ali Farhadi (via Carlos Guestrin and Luke Zettlemoyer) Single-layer neural network 3/9/7 Perceptron as a neural
Improved Local Coordinate Coding using Local Tangents
Improved Local Coordinate Coding using Local Tangents Kai Yu NEC Laboratories America, 10081 N. Wolfe Road, Cupertino, CA 95129 Tong Zhang Rutgers University, 110 Frelinghuysen Road, Piscataway, NJ 08854
Active Learning for Sparse Bayesian Multilabel Classification
Active Learning for Sparse Bayesian Multilabel Classification Deepak Vasisht, MIT & IIT Delhi Andreas Domianou, University of Sheffield Manik Varma, MSR, India Ashish Kapoor, MSR, Redmond Multilabel Classification
Metric Embedding for Kernel Classification Rules
Metric Embedding for Kernel Classification Rules Bharath K. Sriperumbudur University of California, San Diego (Joint work with Omer Lang & Gert Lanckriet) Bharath K. Sriperumbudur (UCSD) Metric Embedding
Introduction to Machine Learning. PCA and Spectral Clustering. Introduction to Machine Learning, Slides: Eran Halperin
1 Introduction to Machine Learning PCA and Spectral Clustering Introduction to Machine Learning, 2013-14 Slides: Eran Halperin Singular Value Decomposition (SVD) The singular value decomposition (SVD)
CSE 417T: Introduction to Machine Learning. Final Review. Henry Chai 12/4/18
CSE 417T: Introduction to Machine Learning Final Review Henry Chai 12/4/18 Overfitting Overfitting is fitting the training data more than is warranted Fitting noise rather than signal 2 Estimating! "#$
PRUNING CONVOLUTIONAL NEURAL NETWORKS. Pavlo Molchanov Stephen Tyree Tero Karras Timo Aila Jan Kautz
PRUNING CONVOLUTIONAL NEURAL NETWORKS Pavlo Molchanov Stephen Tyree Tero Karras Timo Aila Jan Kautz 2017 WHY WE CAN PRUNE CNNS? 2 WHY WE CAN PRUNE CNNS? Optimization failures : Some neurons are "dead":
Semi-supervised Learning
Semi-supervised Learning Introduction Supervised learning: x r, y r R r=1 E.g.x r : image, y r : class labels Semi-supervised learning: x r, y r r=1 R, x u R+U u=r A set of unlabeled data, usually U >>
Asaf Bar Zvi Adi Hayat. Semantic Segmentation
Asaf Bar Zvi Adi Hayat Semantic Segmentation Today s Topics Fully Convolutional Networks (FCN) (CVPR 2015) Conditional Random Fields as Recurrent Neural Networks (ICCV 2015) Gaussian Conditional random
Efficient and Principled Online Classification Algorithms for Lifelon
Efficient and Principled Online Classification Algorithms for Lifelong Learning Toyota Technological Institute at Chicago Chicago, IL USA Talk @ Lifelong Learning for Mobile Robotics Applications Workshop,
A Parallel SGD method with Strong Convergence
A Parallel SGD method with Strong Convergence Dhruv Mahajan Microsoft Research India dhrumaha@microsoft.com S. Sundararajan Microsoft Research India ssrajan@microsoft.com S. Sathiya Keerthi Microsoft Corporation,
Interpreting Deep Classifiers
Ruprecht-Karls-University Heidelberg Faculty of Mathematics and Computer Science Seminar: Explainable Machine Learning Interpreting Deep Classifiers by Visual Distillation of Dark Knowledge Author: Daniela
Lecture 2 - Learning Binary & Multi-class Classifiers from Labelled Training Data
Lecture 2 - Learning Binary & Multi-class Classifiers from Labelled Training Data DD2424 March 23, 2017 Binary classification problem given labelled training data Have labelled training examples? Given
Online Passive-Aggressive Algorithms. Tirgul 11
Online Passive-Aggressive Algorithms Tirgul 11 Multi-Label Classification 2 Multilabel Problem: Example Mapping Apps to smart folders: Assign an installed app to one or more folders Candy Crush Saga 3
Outline. Limits of Bayesian classification Bayesian concept learning Probabilistic models for unsupervised and semi-supervised category learning
Outline Limits of Bayesian classification Bayesian concept learning Probabilistic models for unsupervised and semi-supervised category learning Limitations Is categorization just discrimination among mutually
Machine Learning Basics
Security and Fairness of Deep Learning Machine Learning Basics Anupam Datta CMU Spring 2019 Image Classification Image Classification Image classification pipeline Input: A training set of N images, each
Machine Learning for Signal Processing Bayes Classification and Regression
Machine Learning for Signal Processing Bayes Classification and Regression Instructor: Bhiksha Raj 11755/18797 1 Recap: KNN A very effective and simple way of performing classification Simple model: For
Machine Learning for Structured Prediction
Machine Learning for Structured Prediction Grzegorz Chrupa la National Centre for Language Technology School of Computing Dublin City University NCLT Seminar Grzegorz Chrupa la (DCU) Machine Learning for
Machine Learning for Signal Processing Bayes Classification
Machine Learning for Signal Processing Bayes Classification Class 16. 24 Oct 2017 Instructor: Bhiksha Raj - Abelino Jimenez 11755/18797 1 Recap: KNN A very effective and simple way of performing classification
Kernel Methods and Support Vector Machines
Kernel Methods and Support Vector Machines Oliver Schulte - CMPT 726 Bishop PRML Ch. 6 Support Vector Machines Defining Characteristics Like logistic regression, good for continuous input features, discrete
Being Bayesian About Network Structure:
Being Bayesian About Network Structure: A Bayesian Approach to Structure Discovery in Bayesian Networks Nir Friedman and Daphne Koller Machine Learning, 2003 Presented by XianXing Zhang Duke University
HYPERGRAPH BASED SEMI-SUPERVISED LEARNING ALGORITHMS APPLIED TO SPEECH RECOGNITION PROBLEM: A NOVEL APPROACH
HYPERGRAPH BASED SEMI-SUPERVISED LEARNING ALGORITHMS APPLIED TO SPEECH RECOGNITION PROBLEM: A NOVEL APPROACH Hoang Trang 1, Tran Hoang Loc 1 1 Ho Chi Minh City University of Technology-VNU HCM, Ho Chi
Deep Convolutional Neural Networks for Pairwise Causality
Deep Convolutional Neural Networks for Pairwise Causality Karamjit Singh, Garima Gupta, Lovekesh Vig, Gautam Shroff, and Puneet Agarwal TCS Research, Delhi Tata Consultancy Services Ltd. {karamjit.singh,
Online Videos FERPA. Sign waiver or sit on the sides or in the back. Off camera question time before and after lecture. Questions?
Online Videos FERPA Sign waiver or sit on the sides or in the back Off camera question time before and after lecture Questions? Lecture 1, Slide 1 CS224d Deep NLP Lecture 4: Word Window Classification
Learning Predictive Categories Using Lifted Relational Neural Networks
Learning Predictive Categories Using Lifted Relational Neural Networks Gustav Šourek1, Suresh Manandhar 2, Filip Železný1, Steven Schockaert 3, and Ondřej Kuželka 3 1 Czech Technical University, Prague,
Probabilistic Models for Sequence Labeling
Probabilistic Models for Sequence Labeling Besnik Fetahu June 9, 2011 Besnik Fetahu () Probabilistic Models for Sequence Labeling June 9, 2011 1 / 26 Background & Motivation Problem introduction Generative
Multi-label Active Learning with Auxiliary Learner
Multi-label Active Learning with Auxiliary Learner Chen-Wei Hung and Hsuan-Tien Lin National Taiwan University November 15, 2011 C.-W. Hung & H.-T. Lin (NTU) Multi-label AL w/ Auxiliary Learner 11/15/2011
Expectation propagation as a way of life
Expectation propagation as a way of life Yingzhen Li Department of Engineering Feb. 2014 Yingzhen Li (Department of Engineering) Expectation propagation as a way of life Feb. 2014 1 / 9 Reference This
Lecture 3: Pattern Classification. Pattern classification
EE E68: Speech & Audio Processing & Recognition Lecture 3: Pattern Classification 3 4 5 The problem of classification Linear and nonlinear classifiers Probabilistic classification Gaussians, mitures and
Information Extraction from Text
Information Extraction from Text Jing Jiang Chapter 2 from Mining Text Data (2012) Presented by Andrew Landgraf, September 13, 2013 1 What is Information Extraction? Goal is to discover structured information
Global Optimality in Matrix and Tensor Factorization, Deep Learning & Beyond
Global Optimality in Matrix and Tensor Factorization, Deep Learning & Beyond Ben Haeffele and René Vidal Center for Imaging Science Mathematical Institute for Data Science Johns Hopkins University This
Undirected Graphical Models
Outline Hong Chang Institute of Computing Technology, Chinese Academy of Sciences Machine Learning Methods (Fall 2012) Outline Outline I 1 Introduction 2 Properties Properties 3 Generative vs. Conditional
Text Mining. Dr. Yanjun Li. Associate Professor. Department of Computer and Information Sciences Fordham University
Text Mining Dr. Yanjun Li Associate Professor Department of Computer and Information Sciences Fordham University Outline Introduction: Data Mining Part One: Text Mining Part Two: Preprocessing Text Data
Introduction to Neural Networks
CUONG TUAN NGUYEN SEIJI HOTTA MASAKI NAKAGAWA Tokyo University of Agriculture and Technology Copyright by Nguyen, Hotta and Nakagawa 1 Pattern classification Which category of an input? Example: Character
Principles of Pattern Recognition. C. A. Murthy Machine Intelligence Unit Indian Statistical Institute Kolkata
Principles of Pattern Recognition C. A. Murthy Machine Intelligence Unit Indian Statistical Institute Kolkata e-mail: murthy@isical.ac.in Pattern Recognition Measurement Space > Feature Space >Decision
Global Scene Representations. Tilke Judd
Global Scene Representations Tilke Judd Papers Oliva and Torralba [2001] Fei Fei and Perona [2005] Labzebnik, Schmid and Ponce [2006] Commonalities Goal: Recognize natural scene categories Extract features
Concept Learning. Berlin Chen Department of Computer Science & Information Engineering National Taiwan Normal University.
Concept Learning Berlin Chen Department of Computer Science & Information Engineering National Taiwan Normal University References: 1. Tom M. Mitchell, Machine Learning, Chapter 2 2. Tom M. Mitchell s
Neural Networks. David Rosenberg. July 26, New York University. David Rosenberg (New York University) DS-GA 1003 July 26, / 35
Neural Networks David Rosenberg New York University July 26, 2017 David Rosenberg (New York University) DS-GA 1003 July 26, 2017 1 / 35 Neural Networks Overview Objectives What are neural networks? How
Part-of-Speech Tagging + Neural Networks 3: Word Embeddings CS 287
Part-of-Speech Tagging + Neural Networks 3: Word Embeddings CS 287 Review: Neural Networks One-layer multi-layer perceptron architecture, NN MLP1 (x) = g(xw 1 + b 1 )W 2 + b 2 xw + b; perceptron x is the
Loss Functions and Optimization. Lecture 3-1
Lecture 3: Loss Functions and Optimization Lecture 3-1 Administrative Assignment 1 is released: http://cs231n.github.io/assignments2017/assignment1/ Due Thursday April 20, 11:59pm on Canvas (Extending
Final Exam, Machine Learning, Spring 2009
Name: Andrew ID: Final Exam, 10701 Machine Learning, Spring 2009 - The exam is open-book, open-notes, no electronics other than calculators. - The maximum possible score on this exam is 100. You have 3
Appendix 1 PROOFS 1.1 PROOF OF LEMMA 1
Appendix 1 PROOFS In this section, we provide proofs for all the lemmas and theorems in the main paper. We always assume that a class-conditional extension of the Classification Noise Process (CNP) (Angluin
ECS289: Scalable Machine Learning
ECS289: Scalable Machine Learning Cho-Jui Hsieh UC Davis Oct 27, 2015 Outline One versus all/one versus one Ranking loss for multiclass/multilabel classification Scaling to millions of labels Multiclass
Riemannian Metric Learning for Symmetric Positive Definite Matrices
CMSC 88J: Linear Subspaces and Manifolds for Computer Vision and Machine Learning Riemannian Metric Learning for Symmetric Positive Definite Matrices Raviteja Vemulapalli Guide: Professor David W. Jacobs
Learning from the Wisdom of Crowds by Minimax Entropy. Denny Zhou, John Platt, Sumit Basu and Yi Mao Microsoft Research, Redmond, WA
Learning from the Wisdom of Crowds by Minimax Entropy Denny Zhou, John Platt, Sumit Basu and Yi Mao Microsoft Research, Redmond, WA Outline 1. Introduction 2. Minimax entropy principle 3. Future work and
Active Learning with Cross-class Similarity Transfer
Active Learning with Cross-class Similarity Transfer Yuchen Guo, Guiguang Ding, Yue Gao, Jungong Han Tsinghua National Laboratory for Information Science and Technology (TNList School of Software, Tsinghua
MIRA, SVM, k-nn. Lirong Xia
MIRA, SVM, k-nn Lirong Xia Linear Classifiers (perceptrons) Inputs are feature values Each feature has a weight Sum is the activation activation w If the activation is: Positive: output +1 Negative, output
Support Vector Machine. Industrial AI Lab.
Support Vector Machine Industrial AI Lab. Classification (Linear) Autonomously figure out which category (or class) an unknown item should be categorized into Number of categories / classes Binary: 2 different
Structure Learning in Sequential Data
Structure Learning in Sequential Data Liam Stewart liam@cs.toronto.edu Richard Zemel zemel@cs.toronto.edu 2005.09.19 Motivation. Cau, R. Kuiper, and W.-P. de Roever. Formalising Dijkstra's development
Feature Engineering, Model Evaluations
Feature Engineering, Model Evaluations Giri Iyengar Cornell University gi43@cornell.edu Feb 5, 2018 Giri Iyengar (Cornell Tech) Feature Engineering Feb 5, 2018 1 / 35 Overview 1 ETL 2 Feature Engineering
Apprentissage, réseaux de neurones et modèles graphiques (RCP209) Neural Networks and Deep Learning
Apprentissage, réseaux de neurones et modèles graphiques (RCP209) Neural Networks and Deep Learning Nicolas Thome Prenom.Nom@cnam.fr http://cedric.cnam.fr/vertigo/cours/ml2/ Département Informatique Conservatoire
Lecture 3: Pattern Classification
EE E6820: Speech & Audio Processing & Recognition Lecture 3: Pattern Classification 1 2 3 4 5 The problem of classification Linear and nonlinear classifiers Probabilistic classification Gaussians, mixtures
Performance Comparison of K-Means and Expectation Maximization with Gaussian Mixture Models for Clustering EE6540 Final Project
Performance Comparison of K-Means and Expectation Maximization with Gaussian Mixture Models for Clustering EE6540 Final Project Devin Cornell & Sushruth Sastry May 2015 1 Abstract In this article, we explore
Domain adaptation for deep learning
What you saw is not what you get Domain adaptation for deep learning Kate Saenko Successes of Deep Learning in AI A Learning Advance in Artificial Intelligence Rivals Human Abilities Deep Learning for
Feedback Loop between High Level Semantics and Low Level Vision
Feedback Loop between High Level Semantics and Low Level Vision Varun K. Nagaraja Vlad I. Morariu Larry S. Davis {varun,morariu,lsd}@umiacs.umd.edu University of Maryland, College Park, MD, USA. Abstract.
CENG 793. On Machine Learning and Optimization. Sinan Kalkan
CENG 793 On Machine Learning and Optimization Sinan Kalkan 2 Now Introduction to ML Problem definition Classes of approaches K-NN Support Vector Machines Softmax classification / logistic regression Parzen
Examples are not Enough, Learn to Criticize! Criticism for Interpretability
Examples are not Enough, Learn to Criticize! Criticism for Interpretability Been Kim, Rajiv Khanna, Oluwasanmi Koyejo Wittawat Jitkrittum Gatsby Machine Learning Journal Club 16 Jan 2017 1/20 Summary Examples
Collaborative Filtering. Radek Pelánek
Collaborative Filtering Radek Pelánek 2017 Notes on Lecture the most technical lecture of the course includes some scary looking math, but typically with intuitive interpretation use of standard machine
Laconic: Label Consistency for Image Categorization
1 Laconic: Label Consistency for Image Categorization Samy Bengio, Google with Jeff Dean, Eugene Ie, Dumitru Erhan, Quoc Le, Andrew Rabinovich, Jon Shlens, and Yoram Singer 2 Motivation WHAT IS THE OCCLUDED
Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function.
Bayesian learning: Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function. Let y be the true label and y be the predicted
Learning with Noisy Labels. Kate Niehaus Reading group 11-Feb-2014
Learning with Noisy Labels Kate Niehaus Reading group 11-Feb-2014 Outline Motivations Generative model approach: Lawrence, N. & Scho lkopf, B. Estimating a Kernel Fisher Discriminant in the Presence of
Logistic Regression Logistic
Case Study 1: Estimating Click Probabilities L2 Regularization for Logistic Regression Machine Learning/Statistics for Big Data CSE599C1/STAT592, University of Washington Carlos Guestrin January 10 th,
ECE 661: Homework 10 Fall 2014
ECE 661: Homework 10 Fall 2014 This homework consists of the following two parts: (1) Face recognition with PCA and LDA for dimensionality reduction and the nearest-neighborhood rule for classification;
LINEAR MODELS FOR CLASSIFICATION. J. Elder CSE 6390/PSYC 6225 Computational Modeling of Visual Perception
LINEAR MODELS FOR CLASSIFICATION Classification: Problem Statement 2 In regression, we are modeling the relationship between a continuous input variable x and a continuous target variable t. In classification,
σ(a) = a N (x; 0, 1 2 ) dx. σ(a) = Φ(a) =
Until now we have always worked with likelihoods and prior distributions that were conjugate to each other, allowing the computation of the posterior distribution to be done in closed form. Unfortunately,
Composite Quantization for Approximate Nearest Neighbor Search
Composite Quantization for Approximate Nearest Neighbor Search Jingdong Wang Lead Researcher Microsoft Research http://research.microsoft.com/~jingdw ICML 104, joint work with my interns Ting Zhang from
LOCALITY PRESERVING HASHING. Electrical Engineering and Computer Science University of California, Merced Merced, CA 95344, USA
LOCALITY PRESERVING HASHING Yi-Hsuan Tsai Ming-Hsuan Yang Electrical Engineering and Computer Science University of California, Merced Merced, CA 95344, USA ABSTRACT The spectral hashing algorithm relaxes
Nonparametric Methods Lecture 5
Nonparametric Methods Lecture 5 Jason Corso SUNY at Buffalo 17 Feb. 29 J. Corso (SUNY at Buffalo) Nonparametric Methods Lecture 5 17 Feb. 29 1 / 49 Nonparametric Methods Lecture 5 Overview Previously,
Machine Learning for NLP
Machine Learning for NLP Linear Models Joakim Nivre Uppsala University Department of Linguistics and Philology Slides adapted from Ryan McDonald, Google Research Machine Learning for NLP 1(26) Outline
Lecture 14: Deep Generative Learning
Generative Modeling CSED703R: Deep Learning for Visual Recognition (2017F) Lecture 14: Deep Generative Learning Density estimation Reconstructing probability density function using samples Bohyung Han
Statistical NLP for the Web Log Linear Models, MEMM, Conditional Random Fields
Statistical NLP for the Web Log Linear Models, MEMM, Conditional Random Fields Sameer Maskey Week 13, Nov 28, 2012 1 Announcements Next lecture is the last lecture Wrap up of the semester 2 Final Project
Support Vector Machines. CSE 6363 Machine Learning Vassilis Athitsos Computer Science and Engineering Department University of Texas at Arlington
Support Vector Machines CSE 6363 Machine Learning Vassilis Athitsos Computer Science and Engineering Department University of Texas at Arlington 1 A Linearly Separable Problem Consider the binary classification
ECS289: Scalable Machine Learning
ECS289: Scalable Machine Learning Cho-Jui Hsieh UC Davis Oct 18, 2016 Outline One versus all/one versus one Ranking loss for multiclass/multilabel classification Scaling to millions of labels Multiclass
Neural networks and optimization
Neural networks and optimization Nicolas Le Roux Criteo 18/05/15 Nicolas Le Roux (Criteo) Neural networks and optimization 18/05/15 1 / 85 1 Introduction 2 Deep networks 3 Optimization 4 Convolutional
Machine Learning for NLP
Machine Learning for NLP Uppsala University Department of Linguistics and Philology Slides borrowed from Ryan McDonald, Google Research Machine Learning for NLP 1(50) Introduction Linear Classifiers Classifiers
A Simple Exponential Family Framework for Zero-Shot Learning
A Simple Exponential Family Framework for Zero-Shot Learning Vinay Kumar Verma and Piyush Rai Dept. of Computer Science & Engineering, IIT Kanpur, India {vkverma,piyush}@cse.iitk.ac.in Abstract. We present
FINAL: CS 6375 (Machine Learning) Fall 2014
FINAL: CS 6375 (Machine Learning) Fall 2014 The exam is closed book. You are allowed a one-page cheat sheet. Answer the questions in the spaces provided on the question sheets. If you run out of room for
Generalized Zero-Shot Learning with Deep Calibration Network
Generalized Zero-Shot Learning with Deep Calibration Network Shichen Liu, Mingsheng Long, Jianmin Wang, and Michael I.Jordan School of Software, Tsinghua University, China KLiss, MOE; BNRist; Research
Mirror Descent for Metric Learning. Gautam Kunapuli Jude W. Shavlik
Mirror Descent for Metric Learning Gautam Kunapuli Jude W. Shavlik And what do we have here? We have a metric learning algorithm that uses composite mirror descent (COMID): Unifying framework for metric
Analyzing the Performance of Multilayer Neural Networks for Object Recognition
Analyzing the Performance of Multilayer Neural Networks for Object Recognition Pulkit Agrawal, Ross Girshick, Jitendra Malik {pulkitag,rbg,malik}@eecs.berkeley.edu University of California Berkeley Supplementary
Iterative Laplacian Score for Feature Selection
Iterative Laplacian Score for Feature Selection Linling Zhu, Linsong Miao, and Daoqiang Zhang College of Computer Science and echnology, Nanjing University of Aeronautics and Astronautics, Nanjing 2006,
Kernel Density Topic Models: Visual Topics Without Visual Words
Kernel Density Topic Models: Visual Topics Without Visual Words Konstantinos Rematas K.U. Leuven ESAT-iMinds krematas@esat.kuleuven.be Mario Fritz Max Planck Institute for Informatics mfrtiz@mpi-inf.mpg.de
Large-Margin Thresholded Ensembles for Ordinal Regression
Large-Margin Thresholded Ensembles for Ordinal Regression Hsuan-Tien Lin and Ling Li Learning Systems Group, California Institute of Technology, U.S.A. Conf. on Algorithmic Learning Theory, October 9,
Anticipating Visual Representations from Unlabeled Data. Carl Vondrick, Hamed Pirsiavash, Antonio Torralba
Anticipating Visual Representations from Unlabeled Data Carl Vondrick, Hamed Pirsiavash, Antonio Torralba Overview Problem Key Insight Methods Experiments Problem: Predict future actions and objects Image
Machine Learning Lecture Notes
Machine Learning Lecture Notes Predrag Radivojac January 25, 205 Basic Principles of Parameter Estimation In probabilistic modeling, we are typically presented with a set of observations and the objective
Predictive analysis on Multivariate, Time Series datasets using Shapelets
1 Predictive analysis on Multivariate, Time Series datasets using Shapelets Hemal Thakkar Department of Computer Science, Stanford University hemal@stanford.edu hemal.tt@gmail.com Abstract Multivariate,
Lecture: Face Recognition
Lecture: Face Recognition Juan Carlos Niebles and Ranjay Krishna Stanford Vision and Learning Lab Lecture 12-1 What we will learn today Introduction to face recognition The Eigenfaces Algorithm Linear
Kronecker Decomposition for Image Classification
university of innsbruck institute of computer science intelligent and interactive systems Kronecker Decomposition for Image Classification Sabrina Fontanella 1,2, Antonio Rodríguez-Sánchez 1, Justus Piater