Zeroshot Learning 2015.4.2 Mun Jonghwan.

Slides:



Advertisements
Similar presentations
Query Classification Using Asymmetrical Learning Zheng Zhu Birkbeck College, University of London.
Advertisements

Attributes for Classifier Feedback Amar Parkash and Devi Parikh.
Pseudo-Relevance Feedback For Multimedia Retrieval By Rong Yan, Alexander G. and Rong Jin Mwangi S. Kariuki
Adding Unlabeled Samples to Categories by Learned Attributes Jonghyun Choi Mohammad Rastegari Ali Farhadi Larry S. Davis PPT Modified By Elliot Crowley.
A Unified Framework for Context Assisted Face Clustering
A generic model to compose vision modules for holistic scene understanding Adarsh Kowdle *, Congcong Li *, Ashutosh Saxena, and Tsuhan Chen Cornell University,
Large-Scale Object Recognition using Label Relation Graphs Jia Deng 1,2, Nan Ding 2, Yangqing Jia 2, Andrea Frome 2, Kevin Murphy 2, Samy Bengio 2, Yuan.
Beyond Mindless Labeling: Really Leveraging Humans to Build Intelligent Machines Devi Parikh Virginia Tech.
Data Visualization STAT 890, STAT 442, CM 462
Relative Attributes Presenter: Shuai Zheng (Kyle) Supervised by Philip H.S. Torr Author: Devi Parikh (TTI-Chicago) and Kristen Grauman (UT-Austin)
Generic Object Detection using Feature Maps Oscar Danielsson Stefan Carlsson
Presented by Zeehasham Rasheed
DOG I : an Annotation System for Images of Dog Breeds Antonis Dimas Pyrros Koletsis Euripides Petrakis Intelligent Systems Laboratory Technical University.
Unsupervised Learning of Categories from Sets of Partially Matching Image Features Kristen Grauman and Trevor Darrel CVPR 2006 Presented By Sovan Biswas.
Kuan-Chuan Peng Tsuhan Chen
Bridge Semantic Gap: A Large Scale Concept Ontology for Multimedia (LSCOM) Guo-Jun Qi Beckman Institute University of Illinois at Urbana-Champaign.
Exploiting Ontologies for Automatic Image Annotation M. Srikanth, J. Varner, M. Bowden, D. Moldovan Language Computer Corporation
Review of the web page classification approaches and applications Luu-Ngoc Do Quang-Nhat Vo.
Thanks to Bill Arms, Marti Hearst Documents. Last time Size of information –Continues to grow IR an old field, goes back to the ‘40s IR iterative process.
Universit at Dortmund, LS VIII
Detection, Segmentation and Fine-grained Localization
SYMPOSIUM ON SEMANTICS IN SYSTEMS FOR TEXT PROCESSING September 22-24, Venice, Italy Combining Knowledge-based Methods and Supervised Learning for.
Enhancing Human-Machine Communication via Visual Attributes Devi Parikh Virginia Tech.
What Helps Where – And Why? Semantic Relatedness for Knowledge Transfer Marcus Rohrbach 1,2 Michael Stark 1,2 György Szarvas 1 Iryna Gurevych 1 Bernt Schiele.
Automatic Image Annotation by Using Concept-Sensitive Salient Objects for Image Content Representation Jianping Fan, Yuli Gao, Hangzai Luo, Guangyou Xu.
Semantic Embedding Space for Zero ­ Shot Action Recognition Xun XuTimothy HospedalesShaogang GongAuthors: Computer Vision Group Queen Mary University of.
Beyond Nouns Exploiting Preposition and Comparative adjectives for learning visual classifiers.
C. Lawrence Zitnick Microsoft Research, Redmond Devi Parikh Virginia Tech Bringing Semantics Into Focus Using Visual.
Gang WangDerek HoiemDavid Forsyth. INTRODUCTION APROACH (implement detail) EXPERIMENTS CONCLUSION.
Exploiting Ontologies for Automatic Image Annotation Munirathnam Srikanth, Joshua Varner, Mitchell Bowden, Dan Moldovan Language Computer Corporation SIGIR.
Image Classification over Visual Tree Jianping Fan Dept of Computer Science UNC-Charlotte, NC
Machine Learning Saarland University, SS 2007 Holger Bast Marjan Celikik Kevin Chang Stefan Funke Joachim Giesen Max-Planck-Institut für Informatik Saarbrücken,
26/01/20161Gianluca Demartini Ranking Categories for Faceted Search Gianluca Demartini L3S Research Seminars Hannover, 09 June 2006.
Interactively Discovery of Attributes Vocabulary Devi Parikh and Kristen Grauman.
Object Recognition by Discriminative Combinations of Line Segments and Ellipses Alex Chia ^˚ Susanto Rahardja ^ Deepu Rajan ˚ Maylor Leung ˚ ^ Institute.
Detecting Remote Evolutionary Relationships among Proteins by Large-Scale Semantic Embedding Xu Linhe 14S
Richer Human-Machine Communication in Attributes-based Visual Recognition Devi Parikh TTIC.
Parsing Natural Scenes and Natural Language with Recursive Neural Networks INTERNATIONAL CONFERENCE ON MACHINE LEARNING (ICML 2011) RICHARD SOCHER CLIFF.
Recent developments in object detection
Hybrid Deep Learning for Reflectance Confocal Microscopy Skin Images
CNN-RNN: A Unified Framework for Multi-label Image Classification
Data Driven Attributes for Action Detection
Deep Compositional Cross-modal Learning to Rank via Local-Global Alignment Xinyang Jiang, Fei Wu, Xi Li, Zhou Zhao, Weiming Lu, Siliang Tang, Yueting.
Perceptual Loss Deep Feature Interpolation for Image Content Changes
Mingxia Liu Relative Attributes Mingxia Liu
Perceptrons Lirong Xia.
Part-Based Room Categorization for Household Service Robots
ICCV Hierarchical Part Matching for Fine-Grained Image Classification
Unsupervised Learning and Autoencoders
Using Transductive SVMs for Object Classification in Images
R-CNN region By Ilia Iofedov 11/11/2018 BGU, DNN course 2016.
Thesis Advisor : Prof C.V. Jawahar
Attributes and Simile Classifiers for Face Verification
Attention-based Caption Description Mun Jonghwan.
Rob Fergus Computer Vision
Object Detection + Deep Learning
Deep Visual-Semantic Alignments for Generating Image Descriptions
Word embeddings based mapping
Word embeddings based mapping
RCNN, Fast-RCNN, Faster-RCNN
Using Natural Language Processing to Aid Computer Vision
Adarsh Kowdle*, Congcong Li*, Ashutosh Saxena, and Tsuhan Chen
Meta Learning (Part 2): Gradient Descent as LSTM
Human-object interaction
Presented by: Anurag Paul
Word representations David Kauchak CS158 – Fall 2016.
Motivation It can effectively mine multi-modal knowledge with structured textural and visual relationships from web automatically. We propose BC-DNN method.
Perceptrons Lirong Xia.
Statistical NLP : Lecture 9 Word Sense Disambiguation
Vector Representation of Text
Presentation transcript:

Zeroshot Learning 2015.4.2 Mun Jonghwan

Zero-shot Learning Traindata : Which image shows a cat?

Zero-shot Learning Traindata : Which image shows a giraffe?

Zero-shot Learning Which image shows a giraffe? Description has long neck? Is black? Is spot? lives in plain?

External information Attribute Word vector Hierarchy Co-occurrence - C.H. Lampert, Attribute-based classification for zero-shot visual object classification, TPAMI13[1] D. Parikh, Relative attributes, ICCV2011[3] Z. Akata, Label embedding for attribute-based classification, CVPR13[2] Word vector - A. Frome, Devise: A deep visual-semantic embedding model, NIPS13[3] - Z. Akata, Evaluation of output embedding for fine-grained image classification, CVPR15[4] Hierarchy - Usually used as side information Co-occurrence - T. Mensink, Costa: Co-occurrence statistics for zero-shot classification, CVPR14

Direct Attribute Prediction (DAP)[1] Learn attribute classifier from related classes Use attribute-to-class mapping for prediction Label Attribute Image 𝑝 𝑎 𝑚 = 𝑎 𝑚 𝑧 𝑥 = 𝑝 𝑎 𝑚 𝑥 𝑖𝑓 𝑎 𝑚 𝑧 =1 1−𝑝 𝑎 𝑚 𝑥 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 𝑧 ∗ = argmax 𝑧 𝑚 𝑝( 𝑎 𝑚 𝑧 |𝑥)

Direct Attribute Prediction (DAP)[1] 1. Vocabulary of attributes and class decriptions - giraffe has properties X and Y but not Z 2. Train classifier for each attribute X, Y, Z - From visual examples of related classes 3. Make image attributes predictions 4. Combine into decision: this image is not giraffe 𝑃 𝑋 𝑖𝑚𝑔 =0.8 𝑃 𝑌 𝑖𝑚𝑔 =0.3 𝑃 𝑍 𝑖𝑚𝑔 =0.7 ⇒

Relative Attribute[2] Problem : Binary attributes are very crude If mouse = small, then cat ≠ small If elephant = large, then cat ≠ large 𝑂 𝑚 : 𝑤 𝑚 𝑡 𝑥 𝑖 > 𝑤 𝑚 𝑡 𝑥 𝑗 𝑂 𝑚 : 𝑤 𝑚 𝑡 𝑥 𝑖 = 𝑤 𝑚 𝑡 𝑥 𝑗

Relative Attribute[2] S Clive Smiling J H Age Age: Scarlett Hugh Jared 1 2 ( 𝜇 𝐻 𝑠 + 𝜇 𝑆 𝑠 ) Clive Age: Scarlett Hugh Jared Miley Smiling: Smiling Age S J H 1 2 ( 𝜇 𝐽 𝑠 + 𝑑 𝑚 ) Infer image category using max-likelihood

Attribute Label Embedding (ALE)[2] Embedding to attribute space Search for the class with the highest compatibility

Word Vector[3] Use CNN feature Embedding vector is collected automatically from text corpora Embedding to word vector space

Word Vector[3] Semantically similar classes are close country capital Semantically similar classes are close Word relationship is represented as displacement - 𝐾𝑖𝑛𝑔 – 𝑀𝑎𝑛 + 𝑤𝑜𝑚𝑒𝑛 = 𝑄𝑢𝑒𝑒𝑛

Survey result[4]

Relative information from word vector Tiger : bobcat = strong : ? bobcat : tiger = small : ?

Relative information from word vector Some ranking information Attribute Attribute embedding

Thank you