Search

Your search keyword '"Lifchitz, A."' showing total 323 results
Start Over Author "Lifchitz, A."

Refine your results

more »

more »

more »

more »

more »
323 results on '"Lifchitz, A."'

2. Local Propagation for Few-Shot Learning

Author

Lifchitz, Yann, Avrithis, Yannis, and Picard, Sylvaine

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

The challenge in few-shot learning is that available data is not enough to capture the underlying distribution. To mitigate this, two emerging directions are (a) using local image representations, essentially multiplying the amount of data by a constant factor, and (b) using more unlabeled data, for instance by transductive inference, jointly on a number of queries. In this work, we bring these two ideas together, introducing \emph{local propagation}. We treat local image features as independent examples, we build a graph on them and we use it to propagate both the features themselves and the labels, known and unknown. Interestingly, since there is a number of features per image, even a single query gives rise to transductive inference. As a result, we provide a universally safe choice for few-shot inference under both non-transductive and transductive settings, improving accuracy over corresponding methods. This is in contrast to existing solutions, where one needs to choose the method depending on the quantity of available data., Comment: ICPR 2020

Published
2021

3. Few-Shot Few-Shot Learning and the role of Spatial Attention

Author

Lifchitz, Yann, Avrithis, Yannis, and Picard, Sylvaine

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Few-shot learning is often motivated by the ability of humans to learn new tasks from few examples. However, standard few-shot classification benchmarks assume that the representation is learned on a limited amount of base class data, ignoring the amount of prior knowledge that a human may have accumulated before learning new tasks. At the same time, even if a powerful representation is available, it may happen in some domain that base class data are limited or non-existent. This motivates us to study a problem where the representation is obtained from a classifier pre-trained on a large-scale dataset of a different domain, assuming no access to its training process, while the base class data are limited to few examples per class and their role is to adapt the representation to the domain at hand rather than learn from scratch. We adapt the representation in two stages, namely on the few base class data if available and on the even fewer data of new tasks. In doing so, we obtain from the pre-trained classifier a spatial attention map that allows focusing on objects and suppressing background clutter. This is important in the new problem, because when base class data are few, the network cannot learn where to focus implicitly. We also show that a pre-trained network may be easily adapted to novel classes, without meta-learning.

Published
2020

4. Dense Classification and Implanting for Few-Shot Learning

Author

Lifchitz, Yann, Avrithis, Yannis, Picard, Sylvaine, and Bursuc, Andrei

Subjects
Computer Science - Computer Vision and Pattern Recognition
Abstract

Training deep neural networks from few examples is a highly challenging and key problem for many computer vision tasks. In this context, we are targeting knowledge transfer from a set with abundant data to other sets with few available examples. We propose two simple and effective solutions: (i) dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and (ii) implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features. On miniImageNet, we improve the prior state-of-the-art on few-shot classification, i.e., we achieve 62.5%, 79.8% and 83.8% on 5-way 1-shot, 5-shot and 10-shot settings respectively., Comment: CVPR 2019

Published
2019

7. The genetic basis of a recent transition to live-bearing in marine snails

Author

Stankowski, Sean, Zagrodzka, Zuzanna B., Garlovsky, Martin D., Pal, Arka, Shipilina, Daria, Castillo, Diego Garcia, Lifchitz, Hila, Le Moan, Alan, Leder, Erica, Reeve, James, Johannesson, Kerstin, Westram, Anja M., Butlin, Roger K., Stankowski, Sean, Zagrodzka, Zuzanna B., Garlovsky, Martin D., Pal, Arka, Shipilina, Daria, Castillo, Diego Garcia, Lifchitz, Hila, Le Moan, Alan, Leder, Erica, Reeve, James, Johannesson, Kerstin, Westram, Anja M., and Butlin, Roger K.

Abstract

Key innovations are fundamental to biological diversification, but their genetic basis is poorly understood. A recent transition from egg-laying to live-bearing in marine snails (Littorina spp.) provides the opportunity to study the genetic architecture of an innovation that has evolved repeatedly across animals. Individuals do not cluster by reproductive mode in a genome-wide phylogeny, but local genealogical analysis revealed numerous small genomic regions where all live-bearers carry the same core haplotype. Candidate regions show evidence for live-bearer-specific positive selection and are enriched for genes that are differentially expressed between egg-laying and live-bearing reproductive systems. Ages of selective sweeps suggest that live-bearer-specific alleles accumulated over more than 200,000 generations. Our results suggest that new functions evolve through the recruitment of many alleles rather than in a single evolutionary step.

Published
2024
Full Text
View/download PDF

9. The genetic basis of a recent transition to live-bearing in marine snails

Author

Stankowski, Sean, primary, Zagrodzka, Zuzanna B., additional, Garlovsky, Martin D., additional, Pal, Arka, additional, Shipilina, Daria, additional, Castillo, Diego Garcia, additional, Lifchitz, Hila, additional, Le Moan, Alan, additional, Leder, Erica, additional, Reeve, James, additional, Johannesson, Kerstin, additional, Westram, Anja M., additional, and Butlin, Roger K., additional

Published
2024
Full Text
View/download PDF

10. Contemporary romantics : American piano music.

Author

Lifchitz, Max, 1948- Instrumentalist, Lifchitz, Max, 1948- Elegia., Mazurek, Ron. Voice within., Pizer, Elizabeth, 1954- Expressions intimes., Pizer, Elizabeth, 1954- Strains and restraints., Quilling, Howard, 1935- Sonatas, piano, no. 2., Schiffman, Harold, 1928- Bagatelles, piano., Schiffman, Harold, 1928- Pieces, piano., and Van Appledorn, Mary Jeanne, 1927- Liszt fantasie.

Published
1992

12. Effect of Tuned Parameters on a LSA MCQ Answering Model

Author

Lifchitz, Alain, Jhean-Larose, Sandra, and Denhière, Guy

Subjects
Computer Science - Learning, Computer Science - Artificial Intelligence, Statistics - Machine Learning
Abstract

This paper presents the current state of a work in progress, whose objective is to better understand the effects of factors that significantly influence the performance of Latent Semantic Analysis (LSA). A difficult task, which consists in answering (French) biology Multiple Choice Questions, is used to test the semantic properties of the truncated singular space and to study the relative influence of main parameters. A dedicated software has been designed to fine tune the LSA semantic space for the Multiple Choice Questions task. With optimal parameters, the performances of our simple model are quite surprisingly equal or superior to those of 7th and 8th grades students. This indicates that semantic spaces were quite good despite their low dimensions and the small sizes of training data sets. Besides, we present an original entropy global weighting of answers' terms of each question of the Multiple Choice Questions which was necessary to achieve the model's success., Comment: 9 pages

Published
2008
Full Text
View/download PDF

14. Fast Lexically Constrained Viterbi Algorithm (FLCVA): Simultaneous Optimization of Speed and Memory

Author

Lifchitz, Alain, Maire, Frederic, and Revuz, Dominique

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence, Computer Science - Data Structures and Algorithms, G.2.2, I.5.5, E.2
Abstract

Lexical constraints on the input of speech and on-line handwriting systems improve the performance of such systems. A significant gain in speed can be achieved by integrating in a digraph structure the different Hidden Markov Models (HMM) corresponding to the words of the relevant lexicon. This integration avoids redundant computations by sharing intermediate results between HMM's corresponding to different words of the lexicon. In this paper, we introduce a token passing method to perform simultaneously the computation of the a posteriori probabilities of all the words of the lexicon. The coding scheme that we introduce for the tokens is optimal in the information theory sense. The tokens use the minimum possible number of bits. Overall, we optimize simultaneously the execution speed and the memory requirement of the recognition systems., Comment: 5 pages, 2 figures, 4 tables

Published
2006

15. Ars Nostra plays Max Lifchitz.

Author

Lifchitz, Max, 1948- composer, performer., McCormick, Kim, performer., Falwell, Calvin, performer., Cho, Kyoung, performer., Coble, Jay, performer., Lee, Kisun, performer., McCormick, Robert M., performer., and Lee, Sang-Hie, performer.

Published
2013

17. Diversions.

Author

Lifchitz, Max, 1948- Performer, Alburger, Mark. Twelve fingers., Bulow, Harry. Suite, piano., Deussen, Nancy Bloomer, 1931- Piano music. Selections., Fine, Irving, 1914-1962. Diversions., Woollen, Russell. Berceuse, piano., and Woollen, Russell. Petit rondo.

Published
2002

18. American tapestry : chamber orchestra music by US composers.

Author

Lifchitz, Max, 1948- Conductor, North/South Chamber Orchestra. Performer, Alburger, Mark. Symphonies, no. 1, op. 21,, Jolley, David. Instrumentalist, Kraft, Leo. Unicorno., Meredith, William, oboist. Instrumentalist, Schiffman, Harold, 1928- Concertinos, oboe, chamber orchestra., and Tann, Hilary. Water's edge; arr.

Published
2004

19. Set of five ; Quatrains ; Piano sonata, op. 34, [etc.]

Author

Lifchitz, Max, 1948- Performer, Bell, Larry. Performer, Bell, Larry, 1952- Sonatas, piano, op. 34., Pleskow, Raoul. Quatrains., Quilling, Howard, 1935- Sonatas, piano, No. 4., Rovics, Howard. Sonatas, piano (1987), Toutant, William. Small suite., and Van Appledorn, Mary Jeanne, 1927- Set of five.

Published
1991

20. American debuts : piano music by American composers.

Author

Lifchitz, Max, 1948- Performer, Blank, Allan, 1925- Toccata, piano., Franco, Johan, 1908-1988. Redemption triptych., Franco, Johan, 1908-1988. Theme and variations, piano., Mason, Charles Norman. Blazing macaw., Ovens, Douglas, 1953- Moving image., Satie, Erik, 1866-1925. Morceaux en forme de poire, No. 2; arr., Stewart, Robert, 1918- Variations, piano., and Ziffrin, Marilyn J., 1926- Suite, piano.

Published
1992

29. Exploiter au mieux les données disponibles : représentation et adaptation pour la classification few-shot d'images

Author

Lifchitz, Yann, Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria), Creating and exploiting explicit links between multimedia fragments (LinkMedia), Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-MEDIA ET INTERACTIONS (IRISA-D6), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National de Recherche en Informatique et en Automatique (Inria)-École normale supérieure - Rennes (ENS Rennes)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-CentraleSupélec-IMT Atlantique Bretagne-Pays de la Loire (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-École normale supérieure - Rennes (ENS Rennes)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), Université Rennes 1, Yannis Avrithis, Sylvaine Picard, Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), and Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique)

Subjects
representation learning, Apprentissage profond, Few-Shot, Vision par ordinateur, Représentation, Machine learning, Deep learning, Computer vision, Apprentissage automatique, Adaptation, [INFO.INFO-NE]Computer Science [cs]/Neural and Evolutionary Computing [cs.NE]
Abstract

Deep neural networks can be trained to create highly accurate image classification models, provided we have access to large datasets. In few-shot learning, data is limited to few images, so training from scratch is not feasible. First, a task-independent representation function is learned on abundant data by solving a distinct task such as multiclass classification on a set of base classes. Then, the learned representation is combined with new data of novel classes to solve the few-shot task. In both stages, we introduce solutions that aim at leveraging available data as much as possible. In particular, for representation learning, we propose dense classification training, which for the first time studies local activations in the domain of few-shot learning. We also propose two solutions to adapt the representation function to the few-shot task. Learning is limited to few parameters in implanting or to few gradient updates. Additionally, we study alternative few-shot learning settings, in which access to data is modified. In transductive learning, multiple images need to be classified at the same time. In this context, we propose local propagation, a method that uses similarities between local representations of images to propagate class information. We also introduce few-shot few-shot learning, a new setting, where only few or no in-domain data is accessible for representation learning. In this context, we take advantage of a classifier, pre-trained on a large-scale dataset of a different domain, which can still be adapted to the domain if data is available. In few-shot learning, because data is so scarce, we show that selecting relevant regions with an attention mechanism is important. We propose two simple solutions that successfully fulfill this role. Finally, we apply our knowledge of few-shot learning on the specific problem of classifying aerial images.; Les réseaux de neurones profonds peuvent être entraînés pour produire des modèles de classification d'images très précis, à condition d'avoir accès à un grand nombre de données d'apprentissage. Dans le cas du few-shot learning, les données sont limitées à quelques images ce qui ne permet pas l'apprentissage complet. Dans un premier temps, une fonction de représentation indépendante de la tâche est apprise en résolvant une tâche distincte comme la classification des classes de base. Ensuite, la représentation est combinée avec des exemples des nouvelles classes pour résoudre la tâche few-shot. Pour les deux étapes, nous introduisons des solutions exploitant au mieux les données disponibles. Pour l'apprentissage de représentation, nous proposons la classification dense, qui étudie pour la première fois les activations locales pour le few-shot learning. De plus, nous proposons deux solutions pour adapter la fonction de représentation à la tâche few-shot. L'apprentissage est limité à quelques paramètres dans le cas de l'implantation, ou à quelques itérations. Nous étudions également des problèmes de few-shot learning pour lesquels l'accès à l'information est modifié. Dans le cas du few-shot transductif, plusieurs images doivent être classifiées en même temps. Nous proposons la propagation locale, utilisant les similarités entre représentations locales pour propager l'information de classe. Nous proposons également un nouveau problème, le few-shot few-shot learning, où peu ou aucunes données du domaine n'est accessible. On peut utiliser un réseau pré-entraîné en l'adaptant si possible avec des données du modèle. Pour le few-shot learning, il est important de se focaliser sur les régions pertinentes des images. Nous proposons deux solutions simples d'attention. Enfin, nous appliquons notre savoir dans le cas spécifique de la classification d'images aériennes.

Published
2021

34. Local Propagation for Few-Shot Learning

Author

Yannis Avrithis, Yann Lifchitz, Sylvaine Picard, Creating and exploiting explicit links between multimedia fragments (LinkMedia), MEDIA ET INTERACTIONS (IRISA-D6), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National de Recherche en Informatique et en Automatique (Inria)-École normale supérieure - Rennes (ENS Rennes)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-CentraleSupélec-IMT Atlantique Bretagne-Pays de la Loire (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria), SAFRAN [Paris], SAFRAN Group, Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-MEDIA ET INTERACTIONS (IRISA-D6), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), and Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)

Subjects
FOS: Computer and information sciences, Transduction (machine learning), Distribution (number theory), Computer science, Computer Vision and Pattern Recognition (cs.CV), Shot (filmmaking), Computer Science - Computer Vision and Pattern Recognition, Inference, 02 engineering and technology, 010501 environmental sciences, Machine learning, computer.software_genre, 01 natural sciences, [INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI], Image (mathematics), [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG], Local propagation, 0202 electrical engineering, electronic engineering, information engineering, 0105 earth and related environmental sciences, business.industry, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Contrast (statistics), ComputingMethodologies_PATTERNRECOGNITION, Feature (computer vision), Pattern recognition (psychology), Graph (abstract data type), 020201 artificial intelligence & image processing, Artificial intelligence, business, computer
Abstract

The challenge in few-shot learning is that available data is not enough to capture the underlying distribution. To mitigate this, two emerging directions are (a) using local image representations, essentially multiplying the amount of data by a constant factor, and (b) using more unlabeled data, for instance by transductive inference, jointly on a number of queries. In this work, we bring these two ideas together, introducing \emph{local propagation}. We treat local image features as independent examples, we build a graph on them and we use it to propagate both the features themselves and the labels, known and unknown. Interestingly, since there is a number of features per image, even a single query gives rise to transductive inference. As a result, we provide a universally safe choice for few-shot inference under both non-transductive and transductive settings, improving accuracy over corresponding methods. This is in contrast to existing solutions, where one needs to choose the method depending on the quantity of available data., ICPR 2020

Published
2021
Full Text
View/download PDF

35. Few-Shot Few-Shot Learning and the role of Spatial Attention

Author

Sylvaine Picard, Yannis Avrithis, Yann Lifchitz, Creating and exploiting explicit links between multimedia fragments (LinkMedia), Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-MEDIA ET INTERACTIONS (IRISA-D6), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT), SAFRAN [Paris], SAFRAN Group, MEDIA ET INTERACTIONS (IRISA-D6), Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National de Recherche en Informatique et en Automatique (Inria)-École normale supérieure - Rennes (ENS Rennes)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-CentraleSupélec-IMT Atlantique Bretagne-Pays de la Loire (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Inria Rennes – Bretagne Atlantique, and Institut National de Recherche en Informatique et en Automatique (Inria)

Subjects
FOS: Computer and information sciences, Computer science, business.industry, Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], 02 engineering and technology, 010501 environmental sciences, 01 natural sciences, Classifier (linguistics), 0202 electrical engineering, electronic engineering, information engineering, Task analysis, 020201 artificial intelligence & image processing, Artificial intelligence, business, Classifier (UML), 0105 earth and related environmental sciences
Abstract

International audience; Few-shot learning is often motivated by the ability of humans to learn new tasks from few examples. However, standard few-shot classification benchmarks assume that the representation is learned on a limited amount of base class data, ignoring the amount of prior knowledge that a human may have accumulated before learning new tasks. At the same time, even if a powerful representation is available, it may happen in some domain that base class data are limited or non-existent. This motivates us to study a problem where the representation is obtained from a classifier pre-trained on a large-scale dataset of a different domain, assuming no access to its training process, while the base class data are limited to few examples per class and their role is to adapt the representation to the domain at hand rather than learn from scratch. We adapt the representation in two stages, namely on the few base class data if available and on the even fewer data of new tasks. In doing so, we obtain from the pre-trained classifier a spatial attention map that allows focusing on objects and suppressing background clutter. This is important in the new problem, because when base class data are few, the network cannot learn where to focus implicitly. We also show that a pre-trained network may be easily adapted to novel classes, without meta-learning.

Published
2021

38. O OLHAR DO OUTRO

Author

Miriam Lifchitz Moreira Leite

Subjects
Viajantes, viajantes estrangeiros, século XIX, mulheres brasileiras, viajantes mulheres., History (General), D1-2009
Abstract

Aos estudos de Alfred Schuetz sobre o estrangeiro, o de fora, 0 que não participa do grupo, a década de 90 do século XX assistiu a uma consciência cada vez mais penetrante das distancias entre o nós e os outros. Este artigo limita-se a aplicar alguns dos conhecimentos desenvolvidos ao século XIX, para discutir encontros entre europeus e indígenas e homens e mulheres, no Brasil.

Published
2012

48. Dense Classification and Implanting for Few-Shot Learning

Author

Yannis Avrithis, Andrei Bursuc, Sylvaine Picard, Yann Lifchitz, CentraleSupélec, Creating and exploiting explicit links between multimedia fragments (LinkMedia), MEDIA ET INTERACTIONS (IRISA-D6), Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique Bretagne-Pays de la Loire (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes 1 (UR1), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Institut de Recherche en Informatique et Systèmes Aléatoires (IRISA), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria), SAFRAN Group, VALEO, Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National des Sciences Appliquées (INSA)-Université de Rennes (UNIV-RENNES)-Institut National de Recherche en Informatique et en Automatique (Inria)-École normale supérieure - Rennes (ENS Rennes)-Centre National de la Recherche Scientifique (CNRS)-Université de Rennes 1 (UR1), Université de Rennes (UNIV-RENNES)-CentraleSupélec-IMT Atlantique Bretagne-Pays de la Loire (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Bretagne Sud (UBS)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Inria Rennes – Bretagne Atlantique, Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-MEDIA ET INTERACTIONS (IRISA-D6), Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-Institut National de Recherche en Informatique et en Automatique (Inria)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)-Université de Rennes (UR)-Institut National des Sciences Appliquées - Rennes (INSA Rennes), Institut National des Sciences Appliquées (INSA)-Institut National des Sciences Appliquées (INSA)-Université de Bretagne Sud (UBS)-École normale supérieure - Rennes (ENS Rennes)-CentraleSupélec-Centre National de la Recherche Scientifique (CNRS)-IMT Atlantique (IMT Atlantique), and Institut Mines-Télécom [Paris] (IMT)-Institut Mines-Télécom [Paris] (IMT)

Subjects
FOS: Computer and information sciences, Computer science, Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition, Context (language use), 02 engineering and technology, 010501 environmental sciences, Machine learning, computer.software_genre, 01 natural sciences, Domain (software engineering), Set (abstract data type), 0202 electrical engineering, electronic engineering, information engineering, Feature (machine learning), Layer (object-oriented design), ComputingMilieux_MISCELLANEOUS, 0105 earth and related environmental sciences, business.industry, Deep learning, [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV], Categorization, Metric (mathematics), 020201 artificial intelligence & image processing, Artificial intelligence, business, computer
Abstract

Training deep neural networks from few examples is a highly challenging and key problem for many computer vision tasks. In this context, we are targeting knowledge transfer from a set with abundant data to other sets with few available examples. We propose two simple and effective solutions: (i) dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and (ii) implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features. On miniImageNet, we improve the prior state-of-the-art on few-shot classification, i.e., we achieve 62.5%, 79.8% and 83.8% on 5-way 1-shot, 5-shot and 10-shot settings respectively., CVPR 2019

Published
2019

Catalog

Books, media, physical & digital resources
See catalog results