Machine Learning for Multimedia Material.

Many approaches are used to extract relevant information from multimedia material, ranging from very low-level to higher-level descriptions (classes, captions, ...). That diversity of information is produced by algorithms that have varying degrees of supervision. Lately, fully supervised approaches based on deep learning proved to outperform most older techniques. This is particularly true for the latest developments of Recurrent Neural Networkds (RNN, such as LSTMs) or convolutional neural network (CNNs) for images that reach excellent performance  66. Linkmedia contributes to advancing the state of the art in computing representations for multimedia material by investigating the topics listed below. Some of them go beyond the very processing of multimedia material as they also question the fundamentals of machine learning procedures when applied to multimedia.

• Learning from few samples/weak supervisions. CNNs and RNNs need large collections of carefully annotated data. They are not fitted for analyzing datasets where few examples per category are available or only cheap image-level labels are provided. Linkmedia investigates low-shot, semi-supervised and weakly supervised learning processes: Augmenting scarce training data by automatically propagating labels  69, or transferring what was learned on few very well annotated samples to allow the precise processing of poorly annotated data  78. Note that this context also applies to the processing of heritage collections (paintings, illuminated manuscripts, ...) that strongly differ from contemporary natural images. Not only annotations are scarce, but the learning processes must cope with material departing from what standard CNNs deal with, as classes such as "planes", "cars", etc, are irrelevant in this case.
• Ubiquitous Training. NN (CNNs, LSTMs) are mainstream for producing representations suited for high-quality classification. Their training phase is ubiquitous because the same representations can be used for tasks that go beyond classification, such as retrieval, few-shot, meta- and incremental learning, all boiling down to some form of metric learning. We demonstrated that this ubiquitous training is relatively simpler  69 yet as powerful as ad-hoc strategies fitting specific tasks  83. We study the properties and the limitations of this ubiquitous training by casting metric learning as a classification problem.
• Beyond static learning. Multimedia collections are by nature continuously growing, and ML processes must adapt. It is not conceivable to re-train a full new model at every change, but rather to support continuous training and/or allowing categories to evolve as the time goes by. New classes may be defined from only very few samples, which links this need for dynamicity to the low-shot learning problem discussed here. Furthermore, active learning strategies determining which is the next sample to use to best improve classification must be considered to alleviate the annotation cost and the re-training process  73. Eventually, the learning process may need to manage an extremely large number of classes, up to millions. In this case, there is a unique opportunity of blending the expertise of Linkmedia on large scale indexing and retrieval with deep learning. Base classes can either be "summarized" e.g. as a multi-modal distribution, or their entire training set can be made accessible as an external associative memory  89.
• Learning and lightweight architectures. Multimedia is everywhere, it can be captured and processed on the mobile devices of users. It is necessary to study the design of lightweight ML architectures for mobile and embedded vision applications. Inspired by  93, we study the savings from quantizing hyper-parameters, pruning connections or other approximations, observing the trade-off between the footprint of the learning and the quality of the inference. Once strategy of choice is progressive learning which early aborts when confident enough  74.
• Multimodal embeddings. We pursue pioneering work of Linkmedia on multimodal embedding, i.e., representing multiple modalities or information sources in a single embedded space  87, 86, 88. Two main directions are explored: exploiting adversarial architectures (GANs) for embedding via translation from one modality to another, extending initial work in  88 to highly heterogeneous content; combining and constraining word and RDF graph embeddings to facilitate entity linking and explanation of lexical co-occurrences  63.
• Accountability of ML processes. ML processes achieve excellent results but it is mandatory to verify that accuracy results from having determined an adequate problem representation, and not from being abused by artifacts in the data. Linkmedia designs procedures for at least explaining and possibly interpreting and understanding what the models have learned. We consider heat-maps materializing which input (pixels, words) have the most importance in the decisions  82, Taylor decompositions to observe the individual contributions of each relevance scores or estimating LID 50 as a surrogate for accounting for the smoothness of the space.
• Extracting information. ML is good at extracting features from multimedia material, facilitating subsequent classification, indexing, or mining procedures. Linkmedia designs extraction processes for identifying parts in the images  79, 80, relationships between the various objects that are represented in images  56, learning to localizing objects in images with only weak, image-level supervision  82 or fine-grained semantic information in texts  61. One technique of choice is to rely on generative adversarial networks (GAN) for learning low-level representations. These representations can e.g. be based on the analysis of density  92, shading, albedo, depth, etc.
• Learning representations for time evolving multimedia material. Video and audio are time evolving material, and processing them requests to take their time line into account. In  75, 59 we demonstrated how shapelets can be used to transform time series into time-free high-dimensional vectors, preserving however similarities between time series. Representing time series in a metric space improves clustering, retrieval, indexing, metric learning, semi-supervised learning and many other machine learning related tasks. Research directions include adding localization information to the shapelets, fine-tuning them to best fit the task in which they are used as well as designing hierarchical representations.

Adversarial Machine Learning.

Systems based on ML take more and more decisions on our behalf, and maliciously influencing these decisions by crafting adversarial multimedia material is a potential source of dangers: a small amount of carefully crafted noise imperceptibly added to images corrupts classification and/or recognition. This can naturally impact the insight users get on the multimedia collection they work with, leading to taking erroneous decisions e.g.

This adversarial phenomenon is not particular to deep learning, and can be observed even when using other ML approaches  55. Furthermore, it has been demonstrated that adversarial samples generalize very well across classifiers, architectures, training sets. The reasons explaining why such tiny content modifications succeed in producing severe errors are still not well understood.

We are left with little choice: we must gain a better understanding of the weaknesses of ML processes, and in particular of deep learning. We must understand why attacks are possible as well as discover mechanisms protecting ML against adversarial attacks (with a special emphasis on convolutional neural networks). Some initial contributions have started exploring such research directions, mainly focusing on images and computer vision problems. Very little has been done for understanding adversarial ML from a multimedia perspective  60.

Linkmedia is in a unique position to throw at this problem new perspectives, by experimenting with other modalities, used in isolation one another, as well as experimenting with true multimodal inputs. This is very challenging, and far more complicated and interesting than just observing adversarial ML from a computer vision perspective. No one clearly knows what is at stake with adversarial audio samples, adversarial video sequences, adversarial ASR, adversarial NLP, adversarial OCR, all this being often part of a sophisticated multimedia processing pipeline.

Our ambition is to lead the way for initiating investigations where the full diversity of modalities we are used to work with in multimedia are considered from a perspective of adversarial attacks and defenses, both at learning and test time. In addition to what is described above, and in order to trust the multimedia material we analyze and/or the algorithms that are at play, Linkmedia investigates the following topics:

• Beyond classification. Most contributions in relation with adversarial ML focus on classification tasks. We started investigating the impact of adversarial techniques on more diverse tasks such as retrieval  49. This problem is related to the very nature of euclidean spaces where distances and neighborhoods can all be altered. Designing defensive mechanisms is a natural companion work.
• Detecting false information. We carry-on with earlier pioneering work of Linkmedia on false information detection in social media. Unlike traditional approaches in image forensics  64, we build on our expertise in content-based information retrieval to take advantage of the contextual information available in databases or on the web to identify out-of-context use of text or images which contributed to creating a false information  76.
• Deep fakes. Progress in deep ML and GANs allow systems to generate realistic images and are able to craft audio and video of existing people saying or doing things they never said or did 72. Gaining in sophistication, these machine learning-based "deep fakes" will eventually be almost indistinguishable from real documents, making their detection/rebutting very hard. Linkmedia develops deep learning based counter-measures to identify such modern forgeries. We also carry on with making use of external data in a provenance filtering perspective  81 in order to debunk such deep fakes.
• Distributions, frontiers, smoothness, outliers. Many factors that can possibly explain the adversarial nature of some samples are in relation with their distribution in space which strongly differs from the distribution of natural, genuine, non adversarial samples. We are investigating the use of various information theoretical tools that facilitate observing distributions, how they differ, how far adversarial samples are from benign manifolds, how smooth is the feature space, etc. In addition, we are designing original adversarial attacks and develop detection and curating mechanisms  50.

Multimedia Knowledge Extraction.

Information obtained from collections via computer ran processes is not the only thing that needs to be represented. Humans are in the loop, and they gradually improve their level of understanding of the content and nature of the multimedia collection. Discovering knowledge and getting insight is involving multiple people across a long period of time, and what each understands, concludes and discovers must be recorded and made available to others. Collaboratively inspecting collections is crucial. Ontologies are an often preferred mechanism for modeling what is inside a collection, but this is probably limitative and narrow.

Linkmedia is concerned with making use of existing strategies in relation with ontologies and knowledge bases. In addition, Linkmedia uses mechanisms allowing to materialize the knowledge gradually acquired by humans and that might be subsequently used either by other humans or by computers in order to better and more precisely analyze collections. This line of work is instantiated at the core of the iCODA project Linkmedia coordinates.

We are therefore concerned with:

• Multimedia analysis and ontologies. We develop approaches for linking multimedia content to entities in ontologies for text and images, building on results in multimodal embedding to cast entity linking into a nearest neighbor search problem in a high-dimensional joint embedding of content and entities  86. We also investigate the use of ontological knowledge to facilitate information extraction from content  63.
• Explainability and accountability in information extraction. In relation with ontologies and entity linking, we develop innovative approaches to explain statistical relations found in data, in particular lexical or entity co-occurrences in textual data, for example using embeddings constrained with translation properties of RDF knowledge or path-based explanation within RDF graphs. We also work on confidence measures in entity linking and information extraction, studying how the notions of confidence and information source can be accounted for in knowledge basis and used in human-centric collaborative exploration of collections.
• Dynamic evolution of models for information extraction. In interactive exploration and information extraction, e.g., on cultural or educational material, knowledge progressively evolves as the process goes on, requiring on-the-fly design of new models for content-based information extractors from very few examples, as well as continuous adaptation of the models. Combining in a seamless way low-shot, active and incremental learning techniques is a key issue that we investigate to enable this dynamic mechanisms on selected applications.

Searching.

Search engines must run similarity searches very efficiently. High-dimensional indexing techniques therefore play a central role. Yet, recent contributions in ML suggest to revisit indexing in order to adapt to the specific properties of modern features describing contents.

• Advanced scalable indexing. High-dimensional indexing is one of the foundations of Linkmedia. Modern features extracted from the multimedia material with the most recent ML techniques shall be indexed as well. This, however, poses a series of difficulties due to the dimensionality of these features, their possible sparsity, the complex metrics in use, the task in which they are involved (instance search, $k$-nn, class prototype identification, manifold search  69, time series retrieval, ...). Furthermore, truly large datasets require involving sketching  53, secondary storage and/or distribution  52, 51, alleviating the explosion of the number of features to consider due to their local nature or other innovative methods  68, all introducing complexities. Last, indexing multimodal embedded spaces poses a new series of challenges.
• Improving quality. Scalable indexing techniques are approximate, and what they return typically includes a fair amount of false positives. Linkmedia works on improving the quality of the results returned by indexing techniques. Approaches taking into account neighborhoods  62, manifold structures instead of pure distance based similarities  69 must be extended to cope with advanced indexing in order to enhance quality. This includes feature selection based on intrinsic dimensionality estimation  50.
• Dynamic indexing. Feature collections grow, and it is not an option to fully reindex from scratch an updated collection. This trivially applies to the features directly extracted from the media items, but also to the base class prototypes that can evolve due to the non-static nature of learning processes. Linkmedia will continue investigating what is at stake when designing dynamic indexing strategies.

Navigating.

Navigating a multimedia collection is very central to its understanding. It differs from searching as navigation is not driven by any specific query. Rather, it is mostly driven by the relationships that various documents have one another. Relationships are supported by the links between documents and/or parts of documents. Links rely on semantic similarity, depicting the fact that two documents share information on the same topic. But other aspects than semantics are also at stake, e.g., time with the dates of creation of the documents or geography with mentions or appearance in documents of some geographical landmarks or with geo-tagged data.

In multimedia collections, links can be either implicit or explicit, the latter being much easier to use for navigation. An example of an implicit link can be the name of someone existing in several different news articles; we, as humans, create a mental link between them. In some cases, the computer misses such configurations, leaving such links implicit. Implicit links are subject to human interpretation, hence they are sometimes hard to identify for any automatic analysis process. Implicit links not being materialized, they can therefore hardly be used for navigation or faceted search. Explicit links can typically be seen as hyperlinks, established either by content providers or, more aligned with Linkmedia, automatically determined from content analysis. Entity linking (linking content to an entity referenced in a knowledge base) is a good example of the creation of explicit links. Semantic similarity links, as investigated in the LIMAH project and as considered in the search and hyperlinking task at MediaEval and TRECVid, are also prototypical links that can be made explicit for navigation. Pursuing work, we investigate two main issues:

• Improving multimodal content-based linking. We exploit achievements in entity linking to go beyond lexical or lexico-visual similarity and to provide semantic links that are easy to interpret for humans; carrying on, we work on link characterization, in search of mechanisms addressing link explainability (i.e., what is the nature of the link), for instance using attention models so as to focus on the common parts of two documents or using natural language generation; a final topic that we address is that of linking textual content to external data sources in the field of journalism, e.g., leveraging topic models and cue phrases along with a short description of the external sources.
• Dynamicity and user-adaptation. One difficulty for explicit link creation is that links are often suited for one particular usage but not for another, thus requiring creating new links for each intended use; whereas link creation cannot be done online because of its computational cost, the alternative is to generate (almost) all possible links and provide users with selection mechanisms enabling personalization and user-adaptation in the exploration process; we design such strategies and investigate their impact on exploration tasks in search of a good trade-off between performance (few high-quality links) and genericity.

Summarizing.

Multimedia collections contain far too much information to allow any easy comprehension. It is mandatory to have facilities to aggregate and summarize a large body on information into a compact, concise and meaningful representation facilitating getting insight. Current technology suggests that multimedia content aggregation and story-telling are two complementary ways to provide users with such higher-level views. Yet, very few studies already investigated these issues. Recently, video or image captioning  90, 85 have been seen as a way to summarize visual content, opening the door to state-of-the-art multi-document text summarization  65 with text as a pivot modality. Automatic story-telling has been addressed for highly specific types of content, namely TV series  57 and news  77, 84, but still need a leap forward to be mostly automated, e.g., using constraint-based approaches for summarization  54, 84.

Furthermore, not only the original multimedia material has to be summarized, but the knowledge acquired from its analysis is also to summarize. It is important to be able to produce high-level views of the relationships between documents, emphasizing some structural distinguishing qualities. Graphs establishing such relationships need to be constructed at various level of granularity, providing some support for summarizing structural traits.

Summarizing multimedia information poses several scientific challenges that are:

• Choosing the most relevant multimedia aggregation type: Taking a multimedia collection into account, a same piece of information can be present in several modalities. The issue of selecting the most suitable one to express a given concept has thus to be considered together with the way to mix the various modalities into an acceptable production. Standard summarization algorithms have to be revisited so that they can handle continuous representation spaces, allowing them to benefit from the various modalities  58.
• Expressing user’s preferences: Different users may appreciate quite different forms of multimedia summaries, and convenient ways to express their preferences have to be proposed. We for example focus on the opportunities offered by the constraint-based framework.
• Evaluating multimedia summaries: Finding criteria to characterize what a good summary is remains challenging, e.g., how to measure the global relevance of a multimodal summary and how to compare information between and across two modalities. We tackle this issue particularly via a collaboration with A. Smeaton at DCU, comparing the automatic measures we will develop to human judgments obtained by crowd-sourcing.
• Taking into account structuring and dynamicity: Typed links between multimedia fragments, and hierarchical topical structures of documents obtained via work previously developed within the team are two types of knowledge which have seldom been considered as long as summarization is concerned. Knowing that the event present in a document is causally related to another event described in another document can however modify the ways summarization algorithms have to consider information. Moreover the question of producing coarse-to-fine grain summaries exploiting the topical structure of documents is still an open issue. Summarizing dynamic collections is also challenging and it is one of the questions we consider.

Social biases in text generation.

Recent advances in the domain of text generation allow realistic text-based interaction with a computer. These systems rely on complex neural architectures that leverage very large amount of training texts collected the Web. The problem is that these texts contains unwanted biases (sexism, racism, harmful language...) that are sometimes even amplified by the training procedure. Curating the training texts once for all is not feasible due to the complexity of defining a priori what is relevant or not at the training time. Our work on controlled generation 35, 43 takes another point of view and tries to impose constraints at the inference time. This work aims at making the text generation respect application-specific conditions with the help of a simple classifier.

7.1.1 SurFree

• Name:
  A fast surrogate-free black-box attack against classifier
• Keywords:
  Computer vision, Classification, Cyber attack
• Scientific Description:

  Machine learning classifiers are critically prone to evasion attacks. Adversarial examples are slightly modified inputs that are then misclassified, while remaining perceptively close to their originals. Last couple of years have witnessed a striking decrease in the amount of queries a black box attack submits to the target classifier, in order to forge adversarials. This particularly concerns the blackbox score-based setup, where the attacker has access to top predicted probabilites: the amount of queries went from to millions of to less than a thousand.

  This paper presents SurFree, a geometrical approach that achieves a similar drastic reduction in the amount of queries in the hardest setup: black box decision-based attacks (only the top-1 label is available). We first highlight that the most recent attacks in that setup, HSJA, QEBA and GeoDA all perform costly gradient surrogate estimations. SurFree proposes to bypass these, by instead focusing on careful trials along diverse directions, guided by precise indications of geometrical properties of the classifier decision boundaries. We motivate this geometric approach before performing a head-to-head comparison with previous attacks with the amount of queries as a first class citizen. We exhibit a faster distortion decay under low query amounts (few hundreds to a thousand), while remaining competitive at higher query budgets.

  Paper : https://arxiv.org/abs/2011.12807

• Functional Description:
  This software is the implementation in python of the attack SurFree. This is an attack against a black-box classifier. It finds an input close to the reference input (Euclidean distance) yet not classified with the same predicted label as the reference input. This attack has been tested against image classifier in computer vision.
• URL:
  https://­github.­com/­t-maho/­SurFree
• Authors:
  Thibault Maho, Erwan Le Merrer, Teddy Furon
• Contact:
  Thibault Maho

7.1.2 AML

• Name:
  Asymmetric Metric Learning
• Keywords:
  Knowledge transfer, Metric learning, Image retrieval
• Functional Description:
  This is the official code and a set of pre-trained models that enable the reproduction of the results of our paper https://hal.inria.fr/hal-03047591.
• URL:
  https://­github.­com/­budnikm/­asymmetric_metric_learning
• Contact:
  Ioannis Avrithis

7.1.3 DAL

• Name:
  Rethinking Deep Active Learning
• Keywords:
  Active Learning, Deep learning
• Functional Description:
  This is the official code that enables the reproduction of the results of our paper https://hal.inria.fr/hal-02372102.
• URL:
  https://­github.­com/­osimeoni/­RethinkingDeepActiveLearning
• Contact:
  Ioannis Avrithis

7.1.4 HyperStoryLines

• Keywords:
  Data visualization, Hypergraphs
• Scientific Description:
  HyperStorylines is a technique that generalizes Storylines to visualize the evolution of relationships involving multiple types of entities such as, for example, people, locations, and companies. Datasets which describe such multi-entity relationships are often modeled as hypergraphs, that can be difficult to visualize, especially when these relationships evolve over time. HyperStorylines builds upon Storylines, enabling the aggregation and nesting of these dynamic, multi-entity relationships.
• Functional Description:
  A visualization tool to represent relationships among different types of entities, forming a hypergraph, and that evolve over time.
• URL:
  https://­gitlab.­inria.­fr/­ilda/­hyperstorylines
• Publication:
  hal-03352276
• Contact:
  Vanessa Pena Araya
• Participants:
  Vanessa Pena Araya, Anastasia Bezerianos, Emmanuel Pietriga, Rim Hajri, Laurent Amsaleg

7.1.5 RoBIC

• Name:
  A benchmark suite for assessing classifiers robustness
• Keywords:
  Adversarial attack, Classification, Deep learning
• Functional Description:
  From a dataset of test images annotated with ground truth (like ImageNet), the benchmark leads two attacks to delude the classifier over this set of images and compute some statistics gauging the robustness of the classifier.
• URL:
  https://­gitlab.­inria.­fr/­tmaho/­robustness_benchmark
• Contact:
  Thibault Maho

8.1.1 Training Object Detectors from Few Weakly-Labeled and Many Unlabeled Images

Participants: Zhaohui Yang [Key Laboratory of Machine Perception, Beijing], Miaojing Shi [King‘s College, London], Chao Xu [Key Laboratory of Machine Perception, Beijing], Vittorio Ferrari [Google Research, Zurich], Yannis Avrithis.

Weakly-supervised object detection attempts to limit the amount of supervision by dispensing the need for bounding boxes, but still assumes image-level labels on the entire training set. In this work, we study the problem of training an object detector from one or few images with image-level labels and a larger set of completely unlabeled images 11. This is an extreme case of semi-supervised learning where the labeled data are not enough to bootstrap the learning of a detector. Our solution is to train a weakly-supervised student detector model from image-level pseudo-labels generated on the unlabeled set by a teacher classifier model, bootstrapped by region-level similarities to labeled images. Building upon the recent representative weakly-supervised pipeline PCL, our method can use more unlabeled images to achieve performance competitive or superior to many recent weakly-supervised detection solutions.

8.1.2 Asymmetric metric learning for knowledge transfer

Participants: Mateusz Budnik, Yannis Avrithis.

8.1.3 Local Propagation for Few-Shot Learning

Participants: Yann Lifchitz [Safran, Paris], Yannis Avrithis [Safran, Paris], Sylvaine Picard [Safran, Paris].

8.1.4 Rethinking deep active learning: Using unlabeled data at model training

Participants: Oriane Siméoni [Valeo], Mateusz Budnik, Yannis Avrithis, Guillaume Gravier.

8.1.5 Few-Shot Few-Shot Learning and the role of Spatial Attention

Participants: Yann Lifchitz [Safran, Paris], Yannis Avrithis [Safran, Paris], Sylvaine Picard [Safran, Paris].

8.1.6 Detecting Human-Object Interaction with Mixed Supervision

Participants: Suresh Kumaraswamy [King's college, London], Miaojing Shi [King's college, London], Ewa Kijak.

8.1.7 Attaques et défenses de réseaux de neurones profonds : le cas de la classification d’images

Participants: Hanwei Zhang, Teddy Furon, Laurent Amsaleg, Yannis Avrithis.

8.1.8 Traçage de traîtres

Participant: Teddy Furon.

8.1.9 Detecting Fake News Conspiracies with Multitask and Prompt-Based Learning

Participants: Cheikh Brahim El Vaigh [Shaman], Thomas Girault [Ouest-France], Cyrielle Mallart, Duc Hau Nguyen.

8.1.10 On the hidden treasure of dialog in video question answering

Participants: Deniz Engin [InterDigital Communications], François Schnitzler [InterDigital Communications], Ngoc Duong, Yannis Avrithis.

8.1.11 Iterative label cleaning for transductive and semi-supervised few-shot learning

Participants: Michalis Lazarou, Tania Stathaki, Yannis Avrithis.

8.1.12 Forensics Through Stega Glasses: the Case of Adversarial Images

Participants: Benoît Bonnet [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Teddy Furon [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Patrick Bas [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189].

8.1.13 Generating Adversarial Images in Quantized Domains

Participants: Benoît Bonnet [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Teddy Furon [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Patrick Bas [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189].

8.1.14 SurFree: a fast surrogate-free black-box attack

Participants: Thibault Maho [Wide], Teddy Furon [Wide], Erwan Le Merrer [Wide].

8.1.15 RoBIC: A benchmark suite for assessing classifiers robustness

Participants: Thibault Maho [Wide], Benoît Bonnet [Wide], Teddy Furon [Wide], Erwan Le Merrer [Wide].

8.1.16 Patch Replacement: A Transformation-based Method to Improve Robustness against Adversarial Attacks

Participants: Hanwei Zhang, Yannis Avrithis, Teddy Furon, Laurent Amsaleg.

8.1.17 Efficient Statistical Assessment of Neural Network Corruption Robustness

Participants: Karim Tit [SimSmart], Teddy Furon [SimSmart], Mathias Rousset [SimSmart].

8.1.18 Évaluation statistique efficace de la robustesse de classifieurs

Participants: Karim Tit [SimSmart], Teddy Furon [SimSmart], Mathias Rousset [SimSmart], Louis-Marie Traonouez [Thalès, Rennes].

8.1.19 Probabilistic forecasting of seasonal time series Combining clustering and classification for forecasting

Participants: Colin Leverger [Orange, Rennes], Thomas Guyet [AgroCampus Ouest, Rennes], Simon Malinowski [Orange, Lannion], Vincent Lemaire [Orange, Lannion], Alexis Bondu [Orange, Paris], Laurence Rozé [Lacodam], Alexandre Termier [Lacodam], Régis Marguerie [Orange, Rennes].

8.1.20 Impact of Data Cleansing for Urban Bus Commercial Speed Prediction

Participants: Gauthier Lyan [Diverse], David Gross-Amblard [Druid], Jean-Marc Jézéquel [Diverse], Simon Malinowski.

8.1.21 A survey on training and evaluation of word embeddings

Participants: François Torregrossa [SoLocal, Boulogne-Billancourt], Robin Allesiardo [SoLocal, Boulogne-Billancourt], Vincent Claveau [SoLocal, Boulogne-Billancourt], Nihel Kooli [SoLocal, Boulogne-Billancourt], Guillaume Gravier.

8.1.22 Unsupervised Tree Extraction in Embedding Spaces for Taxonomy Induction

Participants: François Torregrossa [SoLocal, Boulogne-Billancourt], Robin Allesiardo [SoLocal, Boulogne-Billancourt], Vincent Claveau, Guillaume Gravier.

8.1.23 Active Learning for Interactive Relation Extraction in a French Newspaper’s Articles

Participants: Cyrielle Mallart [Sipa Ouest-France, Rennes], Michel Le Nouy [Sipa Ouest-France, Rennes], Guillaume Gravier, Pascale Sébillot.

8.2.1 HyperStorylines: Interactively untangling dynamic hypergraphs – GO TO ACCESSING

Participants: Vanessa Pena Araya [Ilda], Tong Xue [Ilda], Emmanuel Pietriga [Ilda], Laurent Amsaleg [Ilda], Anastasia Bezerianos [Ilda].

8.2.2 Generating artificial texts for query expansion

Participant: Vincent Claveau.

8.2.3 Generating artificial texts as substitution or complement of training data

Participants: Vincent Claveau, Antoine Chaffin, Ewa Kijak.

The quality of artificially generated texts has considerably improved with the advent of transformers. The question of using these models to generate learning data for supervised learning tasks naturally arises. In this series of papers 14, 44, this question is explored under 3 aspects: (i) are artificial data an efficient complement? (ii) can they replace the original data when those are not available or cannot be distributed for confidentiality reasons? (iii) can they improve the explainability of classifiers? Different experiments are carried out on Web-related classification tasks – namely sentiment analysis on product reviews and Fake News detection – using artificially generated data by fine-tuned GPT-2 models. The results show that such artificial data can be used in a certain extend but require pre-processing to significantly improve performance. We show that bag-of-word approaches benefit the most from such data augmentation.

8.2.4 PPL-MCTS: Constrained Textual Generation Through Discriminator-Guided Decoding

Participants: Vincent Claveau, Antoine Chaffin, Ewa Kijak.

Large pre-trained language models (LM) based on Transformers allow to generate very plausible long texts. In this series of papers 35, 43, we explore how this generation can be further controlled to satisfy certain constraints (eg. being non-toxic, positive or negative, convey certain emotions, etc.) without fine-tuning the LM. Precisely, we formalize constrained generation as a tree exploration process guided by a discriminator that indicates how well the associated sequence respects the constraint. Using a discriminator to guide this generation, rather than fine-tuning the LM, in addition to be easier and cheaper to train, allows to apply the constraint more finely and dynamically. We propose several original methods to search this generation tree, notably the Monte Carlo Tree Search (MCTS) which provides theoretical guarantees on the search efficiency, but also simpler methods based on re-ranking a pool of diverse sequences using the discriminator scores. These methods are evaluated on two types of constraints and languages: review polarity and emotion control in French and English. We show that MCTS achieves state-of-the-art results in constrained generation, without having to tune the language model, in both tasks and languages. We also demonstrate that our other proposed methods based on re-ranking can be really effective when diversity among the generated propositions is encouraged.

8.2.5 A Study of the Plausibility of Attention between RNN Encoders in Natural Language Inference

Participants: Duc Hau Nguyen, Guillaume Gravier, Pascale Sébillot.

8.2.6 Understanding the phenomenology of reading through modeling

Participants: Alessio Antonini [OU - The Open University, Milton Keynes], Mari Carmen Suárez-Figueroa [OEG - Ontology Engineering Group, Madrid], Alessandro Adamou [INSIGHT - Insight Centre for Data Analytics, Galway], Francesca Benatti [OU - The Open University, Milton Keynes], François Vignale [3L.AM - Langues, Littératures, Linguistique des universités d'Angers et du Mans], Guillaume Gravier [University of Turin], Lucia Lupi [University of Turin].

8.2.7 Utilisation d'approches automatiques pour la reconnaissance des expériences de lecture–"Je pense que ça traite d'expérience de lecture, à voir ... " : retour sur une expérience d'annotation collaborative

Participants: Guillaume Le Noé Bienvenu [3L.AM - Langues, Littératures, Linguistique des universités d'Angers et du Mans], François Vignale [3L.AM - Langues, Littératures, Linguistique des universités d'Angers et du Mans], Guillaume Gravier, Pascale Sébillot.

Ces communications (21, 33) ont pour but de présenter le rôle des techniques relevant de l'intelligence artificielle et du traitement du langage naturel dans la mise au point d'algorithmes de détection semi-automatique des expériences de lecture développés dans le cadre du projet READ-IT. Au cours des dernières décennies, les connaissances sur l'histoire des pratiques de lecture ont considérablement augmenté au sujet des usages et des habitudes mais des questions fondamentales demeurent, telles que le "pourquoi" et le "comment" on lit. Grâce à l'exploration de sources numériques à la recherche de témoignages d'expériences de lecture, le projet READ-IT (Reading Europe Advanced Data Investigation Tool) vise à mieux comprendre ces phénomènes. Ce projet financé par le Joint Programming Initiative for Cultural Heritage (2018-2021) associe 5 partenaires de 4 pays (France, Royaume-Uni, Pays-Bas, République Tchèque).

CIFRE PhD: Incremental dynamic construction of knowledge bases from text mining

Participants: Guillaume Gravier, Cyrielle Mallart, Pascale Sébillot.

Duration: 3 years, started in Dec. 2018 Partner: Ouest France

CIFRE PhD: Embedding heterogeneous data for directory search

Participants: Vincent Claveau, Guillaume Gravier, François Torregrossa.

Duration: 3 years, started in Dec. 2018 Partner: SoLocal

CIFRE PhD: Few shot learning for object recognition in aerial images

Participants: Yannis Avrithis, Yann Lifchitz.

Duration: 3 years, started in March 2018 Partner: Safran Tech

CIFRE PhD: Deep Learning and Homomorphic encryption

Participants: Teddy Furon, Samuel Tap.

Duration: 3 years, started in December 2020 Partner: ZAMA.ia

CIFRE PhD: Robustness of machine learning against uncertainties

Participants: Teddy Furon, Mathias Rousset, Karim Tit.

Duration: 3 years, started in December 2020 Partner: THALES La Ruche

CIFRE PhD: Semantic multimodal question answering (MQA) in domestic environments

Participants: Yannis Avrithis, Teddy Furon, Deniz Engin.

Duration: 3 years, started in September 2020 Partner: InterDigital

CIFRE PhD: Multimodal detection of fake news

Participants: Vincent Claveau, Ewa Kijak, Antoine Chaffin.

Duration: 3 years, started in November 2020 Partner: IMATAG

10.1.1 Associate Teams in the framework of an Inria International Lab or in the framework of an Inria International Program

10.1.2 Participation in other International Programs

10.2.1 Other european programs/initiatives

Chaire Security of AI for Defense Applications (SAIDA)

Participants: Teddy Furon [WIDE], Laurent Amsaleg [WIDE], Erwan Le Merrer [WIDE], Mathias Rousset [SIMSMART], Benoit Bonnet [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Thibault Maho [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Patrick Bas [CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189], Samuel Tap, Karim Tit.

Duration: 4 years, started Sept 2020 ANR-20-CHIA-0011-01

Inria Project Lab Knowledge-driven data and content collaborative analytics (iCODA)

Participants: Laurent Amsaleg, Cheikh Brahim El Vaigh, Guillaume Gravier, Cyrielle Mallart, Pascale Sébillot.

Duration: 4.5 years, started in April 2017 Partners: Inria project-teams Linkmedia, CEDAR, GraphIK and ILDA, with Ouest-France, Le Monde and AFP

ANR Archival: Multimodal machine comprehension of language for new intelligent interfaces of scientific and cultural mediation

Participants: Laurent Amsaleg, Guillaume Gravier, Duc Hau Nguyen, Pascale Sébillot.

Duration: 3.5 year, started in Dec. 2019

The multidisciplinary and multi-actor ARCHIVAL project aims at yielding collaborations between researchers from the fields of Information and Communication Sciences as well as Computer Sciences around archive value enhancing and knowledge sharing for arts, culture and heritage. The project is structured around the following questionings: What part can machine comprehension methods play towards the reinterpretation of thematic archive collections? How can content mediation interfaces exploit results generated by current AI approaches?

ARCHIVAL teams will explore heterogeneous document collection structuration in order to explicitly reveal implicit links, to explain the nature of these links and to promote them in an intelligible way towards ergonomic mediation interfaces that will guarantee a successful appropriation of contents. A corpus has been delimited from the FMSH “self-management” collection, recently awarded as Collex, which will be completed from the large Canal-U academic audiovisual portal. The analysis and enhancement of this collection is of particular interest for Humanities and Social Sciences in a context where it becomes a necessity to structurally reconsider new models of socioeconomic development (democratic autonomy, social and solidarity-based economy, alternative development,…).

ANR MEERQAT: MultimEdia Entity Representation and Question Answering Tasks

Participants: Laurent Amsaleg, Yannis Avrithis, Ewa Kijak, Shashanka Venkataramanan.

Duration: 3.5 year, started in April 2020 Partners: Inria project-teams Linkmedia, CEA LIST, LIMSI, IRIT.

The overall goal of the project is to tackle the problem of ambiguities of visual and textual content by learning then combining their representations. As a final use case, we propose to solve a Multimedia Question Answering task, that requires to rely on three different sources of information to answer a (textual) question with regard to visual data as well as an external knowledge base containing millions of unique entities, each being represetd by textual and visual content as well as some links to other entities. An important work will deal with the representation of entities into a common tri-modal space, in which one should determine the content to associate to an entity to adequately represent it. The challenge consists in defining a representation that is compact (for performance) while still expressive enough to reflect the potential links between the entity and a variety of others.

MinArm: EVE3

Participants: Teddy Furon.

Duration: 3 year, started in April 2019 Partners: MinArm, CRIStAL Lille, LIRMM, Univ. Troyes, Univ. Paris Saclay

ANR UNLIR: Unsupervised Representation Learning for Image Recognition

Participants: Yannis Avrithis.

Duration: 4 years, started in January 2020In relation with the JCJC awarded to Ronan Sicre, LIS, Aix-Marseille.

1. Complex recognition problems such as fine grained classification (highly similar categories e.g. bird species, airplane/car models, etc.) show that state of the art DNNs are still improved by better objective functions and more discriminative intermediate representations.
2. Despite progress in using less annotated data, DNN can hardly cope with learning from few examples.
3. DNNs have so many parameters and complex structures that it is extremely hard to understand what happens in every layer in producing the final decision.

This project aims to address these limitations. In particular, we will work towards building networks capable of solving fine-grained visual recognition tasks. We will improve the capabilities of networks to learn from few to no data, building highly discriminative representations that can address complex recognition problems. Following that, we will provide insight on how such models take their decisions.

AID-CNRS: FakeNews

Participants: Vincent Claveau, Ewa Kijak.

Duration: 2 years, started mid-2021

11.1.1 Scientific events: organisation

11.1.2 Scientific events: selection

11.1.3 Journal

11.1.4 Invited talks

• Teddy Furon, HubDay Data Science & Artificial Intelligence (Pôle de compétitivité Systematic)
• Teddy Furon, UK/FR Defense Sector Trusted AI Workshop organized by DGA/AID and MOD/DSTL
• Teddy Furon, Journées Scientifiques Inria
• Teddy Furon, Journées Nationales du GdR Securité Informatique
• Teddy Furon, Journée annuelle du programme Confiance.ai (IRT SystemX), atelier robustesse de l'IA
• Teddy Furon, Atelier Hands-on Machine Learning for Security, CentraleSupelec Rennes
• Teddy Furon, European CyberWeek, Présentation du Centre de Compétence en Cybersécurité C3
• Teddy Furon, IEEE Workshop on Information Forensics ans Security (keynote speaker)
• Teddy Furon, Internal seminars (Thales, Amazon, Content Armor, Insa Rennes lab VADDER)
• Vincent Claveau presented LinkMedia's work on fake news detection at 2-day multidisciplinary event of Univ. Avignon
• Vincent Claveau presented LinkMedia's work on fake news detection at the IXXI seminar
• Guillaume Gravier, L'IA pour décrire, indexer et explorer des collections multimédias : comment faire sans Google ? Conférence invité d'ouverture du colloque Humanistica
• Ewa Kijak, Manipulation des images, des vidéos et fausses informations : qu’est-ce que l’IA a changé ?, Breizh Data days
• Ewa Kijak, Fake news, Deepfakes : un défi pour la recherche en informatique, Dakar Institute of Technology
• Ewa Kijak, Deep Learning Breakthrough in Computer Vision, Dakar Institute of Technology, Percepts and Concepts seminar, ENS ULM

11.1.5 Leadership within the scientific community

• Laurent Amsaleg is a member of the Steering Committee of ACM Multimedia for the 2020-2023 term
• Pascale Sébillot is a member of the board of the GDR Traitement Automatique des Langues
• Guillaume Gravier is a member of the scientific board of the GDR Traitement Automatique des Langues
• Guillaume Gravier is a referent on AI within Allistene representing the CPU
• Guillaume Gravier is the coordinator of Rennes' AI doctoral program AI4SDA

11.1.6 Scientific expertise

• Teddy Furon was an expert for Pôle d'Excellence en Cybersécurité, IDEX Grenoble CBS-MIAI.
• Teddy Furon was a member of committee ANR ASTRID AAP.
• Vincent Claveau served as expert for ANRT (CIFRE PhD funding)
• Vincent Claveau served as expert for LabEx Digicosme
• Vincent Claveau served as jury member for the AI4SDA PhD funding program

11.1.7 Research administration

• Pascale Sébillot is a deputy director of IRISA UMR 6074 since January 1st, 2021
• Pascale Sébillot is the deputy director of the Scientific Advisory Committee of IRISA UMR 6074
• Pascale Sébillot was a member of the theses advisory committee of the MathSTIC doctoral school until April 2021
• Pascale Sébillot is a member of the board of the MathSTIC doctoral school
• Teddy Furon is a member of Commission du Personnel of IRISA / Inria Rennes Bretagne Atlantique
• Teddy Furon was a member of Commission des Moyens Incitatifs for Inria Rennes Bretagne Atlantique
• Teddy Furon was the head of the Comité de sélection des délégations at Inria Rennes Bretagne Atlantique
• Guillaume Gravier is director of IRISA (UMR 6074) since Jan. 2021
• Guillaume Gravier is a member of the board of the competitvity cluster Images & Réseaux, representing CNRS
• Guillaume Gravier is a member of the AID-CNRS coordination on AI

11.2.1 Teaching

• Licence: Pascale Sébillot, Natural Language Processing, 9h, L3, INSA Rennes, France
• Master: Laurent Amsaleg, Bases de données avancées, 25h, M2, INSA Rennes, France
• Master: Teddy Furon, Rare Event Simulations, 40h, INSA Rennes, France
• Master: Pascale Sébillot, Natural Language Processing, 6h, M1, INSA Rennes, France
• Engineering school: Vincent Claveau, Machine Learning, 18h, 3rd year, INSA Rennes, France
• Master: Vincent Claveau, Information Retrieval, 10h, M2 MIAGE, Univ. Rennes, France
• Engineering school: Vincent Claveau, Natural Language Processing, 12h, ESIR, Univ. Rennes, France
• Licence: Guillaume Gravier, Base de données, 26h, L2, INSA Rennes
• Licence: Guillaume Gravier, Natural language processing, 12h, L3, INSA Rennes
• Licence: Guillaume Gravier, Markov models, 6h, L3, INSA Rennes
• Master: Guillaume Gravier, Natural Language Processing, 6h, M1, INSA Rennes
• Master: Guillaume Gravier, Natural Language Processing, 33h, M2, ENSAI
• Master: Ewa Kijak, Image processing, 45h, M1, ESIR, Univ. Rennes
• Master: Ewa Kijak, Supervised machine learning, 15h, M2R, Univ. Rennes
• Master: Ewa Kijak, Image classification, 27h, M1, ESIR
• Ewa Kijak is head of the Image engineering track (M1-M2) of ESIR, Univ. Rennes
• Master: Ewa Kijak, Image classification, 27h, M1, ESIR
• Master: Yannis Avrithis, Deep learning for vision, 20h, M2 SIF France
• Master: Simon Malinowski, Basics of Data Analytics for Data Science, 24h, EIT Data Science Master 1, Rennes
• Master: Simon Malinowski, Prediction Methods, 30h, M1 MIAGE and Data Science EIT Master 1, Rennes
• Master: Simon Malinowski, Statisical Data Mining, 24h, M2 MIAGE, ISTIC, Rennes
• Master: Simon Malinowski, Symbolic Data Mining, 12h, M2 MIAGE, ISTIC, Rennes
• Simon Malinowski is responsible for the Master 2 MIAGE parcours Classique
• Simon Malinowski is responsible for the M2 studies within the DataScience track of the EIT-digital master school

11.2.2 Supervision

• PhD in progress: Shashanka Venkataramanan, Metric learning for instance- and category-level visual representations. Started in December 2020. Yannis Avrithis, Ewa Kijak & Laurent Amsaleg
• PhD in progress: Benoit Bonnet, Security of DNN against white-box attacks. Started in November 2019. Teddy Furon & Patrick Bas (CNRS CRIStAL Lille)
• PhD in progress: Thibault Maho, Security of DNN against black-box attacks. Started in December 2020. Erwan Le Merrer & Teddy Furon
• PhD in progress: Karim Tit, Assessment of DNN robustness against uncertainties. Started in February 2021. Teddy Furon
• PhD in progress: Samuel Tap, Homomorphic encryption for Machine Learning. Started in March 2021. Teddy Furon
• PhD in progress: Deniz Engin, Video Query Answering in domestic environments. Started in September 2020. Yannis Avrithis & Teddy Furon
• PhD in progress: Cyrielle Mallart, Incremental dynamic construction of knowledge graphs from text mining, started December 2018, Guillaume Gravier, Michel Le Nouy (Ouest-France)& Pascale Sébillot
• PhD in progress: Duc Hau Nguyen, Multimodal space for the generation and justification of semantic links between documents, started September 2020, Guillaume Gravier & Pascale Sébillot
• PhD in progress: Antoine Chaffin, Multimodal Fake News Detection, started in October 2020, Ewa Kijak & Vincent Claveau
• PhD in progress: Mohamed Younes, Learning and simulating strategies in sports for VR training, started in December 2020, Franck Multon (Mimetic), Richard Kulpa (Mimetic), Ewa Kijak, Simon Malinowski
• PhD: Yichang Wang, Interpretable time series classification, defended September 2021, Élisa Fromont (Lacodam), Romain Tavenard (Obélix), Rémi Emonet (Jean Monnet U.), Simon Malinowski
• PhD: Cheikh Brahim El Vaigh, Incremental content to data linking leveraging ontological knowledge in data journalism 39, defended January 2021, Guillaume Gravier, Pascale Sébillot & François Goasdoué (with CEDAR, Inria team)
• PhD: Hanwei Zhang, Deep Learning in Adversarial Contexts 42, defended June 2021, Laurent Amsaleg, Yannis Avrithis & Teddy Furon
• PhD: Marzieh Gheisari Khorasgani, Secure identification in the Internet of Things 40, defended July 2021, Laurent Amsaleg & Teddy Furon
• PhD: Yann Lifchitz, Making the most of available data : representation and adaptation for few-shot image classification 41, defended April 2021, Yannis Avrithis & Sylvaine Picard (Safran)
• PhD: François Torregrossa, Hyperbolic embeddings for the represenation of words and knowledge, defended Dec. 2021, Guillaume Gravier & Vincent Claveau

11.2.3 Juries

• Laurent Amsaleg was a reviewer for Solène Bernard's PhD committee (Centrale Lille. Oct 2021)
• Teddy Furon was a reviewer for Alexandre Araujo's PhD committee (Univ. Paris Dauphine P.S.L. June 2021)
• Teddy Furon was a reviewer for Rémi Bernhard's PhD committee (Mines de St Etienne. Nov 2021)
• Teddy Furon was a member of Solène Bernard's PhD committee (Centrale Lille. Oct 2021)
• Vincent Claveau was a reviewer for Perceval Wajsburt's PhD committee (Sorbonne Université)
• Vincent Claveau was a member of Marwa Elleuch'PhD committee (Institut Polytechnique de Paris)
• Pascale Sébillot was president of Antoine Perquin's PhD committee (Université de Rennes 1).
• Pascale Sébillot was a reviewer of Aman Behre's PhD committee (Université Paris-Saclay)
• Pascale Sébillot was a reviewer of Michael Filhol's habilitation (HDR) committee (Université Paris-Saclay)
• Guillaume Gravier was reviewer for Sebastian Ferrada's PhD (UChile, Santiago de Chile)
• Guillaume Gravier was president of the jury of Hugo Talibart's PhD (Université de Rennes 1)
• Ewa Kijak was a member of Florent Guiotte's PhD committee (Université de Rennes 2)
• Ewa Kijak was a member of Sanae Bourtafass's PhD committee (Université de La Rochelle)

11.3.1 Internal or external Inria responsibilities

Linkmedia was representing the lab IRISA and the Inria research center at "La Fête de la Science" during a week in October 2021 at Les Champs Libres, Rennes.

Vincent Claveau was in charge of organizing the first popularization event of the GdR TAL about recent advances in NLP.

11.3.2 Articles and contents

Rennes – une IA souveraine au service de la vie publique 47, Contribution au Bulletin de l'Association Française pour l'Intelligence Artificielle, AFIA.

Participants: Guillaume Gravier [Lacodam], Elisa Fromont [Lacodam], Nicolas Courty [Obélix], Teddy Furon [Sirocco], Christine Guillemot [Sirocco], Paolo Robuffo Giordano [Rainbow].

11.3.3 Interventions

Vincent Claveau was interviewed about fake news and deepfake detection by the RTS and Data Analytics Post

Ewa Kijak was interviewed about IA in fake news and deepfake detection by the RFI, Data Analytics Post, and Science et vie

Guillaume Gravier gave a webinar on AI for the competivity cluster Images & Réseaux

Laurent Amsaleg participated to two panels during the "DigitalTech" series of conferences about Fake News and Deep Fakes.

International journals

• 7 articleA.Alessio Antonini, M.Mari Carmen Suárez-Figueroa, A.Alessandro Adamou, F.Francesca Benatti, F.François Vignale, G.Guillaume Gravier and L.Lucia Lupi. Understanding the phenomenology of reading through modelling.Semantic Web – Interoperability, Usability, Applicability122January 2021, 191-217
  HALDOIback to text
• 8 articleG.Gauthier Lyan, D.David Gross-Amblard, J.-M.Jean-Marc Jézéquel and S.Simon Malinowski. Impact of Data Cleansing for Urban Bus Commercial Speed Prediction.Springer Nature Computer ScienceNovember 2021, 1-11
  HALback to text
• 9 articleV.Vanessa Pena Araya, T.Tong Xue, E.Emmanuel Pietriga, L.Laurent Amsaleg and A.Anastasia Bezerianos. HyperStorylines: Interactively untangling dynamic hypergraphs.Information VisualizationSeptember 2021, 1-21
  HALDOIback to text
• 10 articleF.François Torregrossa, R.Robin Allesiardo, V.Vincent Claveau, N.Nihel Kooli and G.Guillaume Gravier. A survey on training and evaluation of word embeddings.International Journal of Data Science and Analytics112February 2021, 85-103
  HALDOIback to text
• 11 articleZ.Zhaohui Yang, M.Miaojing Shi, C.Chao Xu, V.Vittorio Ferrari and Y.Yannis Avrithis. Training Object Detectors from Few Weakly-Labeled and Many Unlabeled Images.Pattern Recognition120December 2021, 108164
  HALDOIback to text

International peer-reviewed conferences

• 12 inproceedingsB.Benoît Bonnet, T.Teddy Furon and P.Patrick Bas. Forensics Through Stega Glasses: the Case of Adversarial Images.ICPR-MMForWILD 2020 - Workshop MultiMedia FORensics in the WILDPattern Recognition: ICPR International Workshops and ChallengesMilan, ItalySpringer International PublishingJanuary 2021, 1-17
  HALback to text
• 13 inproceedingsM.Mateusz Budnik and Y.Yannis Avrithis. Asymmetric metric learning for knowledge transfer.CVPR 2021 - IEEE/CVF Conference on Computer Vision and Pattern RecognitionVirtual, FranceIEEEJune 2021, 8224-8234
  HALDOIback to text
• 14 inproceedingsV.Vincent Claveau, A.Antoine Chaffin and E.Ewa Kijak. La génération de textes artificiels en substitution ou en complément de données d’apprentissage.Actes de la 28e Conférence sur le Traitement Automatique des Langues Naturelles : conférence principaleTALN 2021 - 28e Conférence sur le Traitement Automatique des Langues Naturelles1Lille, FranceATALA2021, 37-49
  HALback to text
• 15 inproceedingsV.Vincent Claveau. Generating artificial texts for query expansion.Actes de la conférence CORIA 2021CORIA 2021 - Conférence en Recherche d’Information et ApplicationsGrenoble, FranceApril 2021, 1-16
  HALback to text
• 16 inproceedingsV.Vincent Claveau. Neural text generation for query expansion in information retrieval.WI-IAT 2021 - 20th IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent TechnologyProceedings of the WI-IAT ConferenceMelbourne, AustraliaIEEEDecember 2021, 1-8
  HALDOIback to text
• 17 inproceedingsC. B.Cheikh Brahim El Vaigh, T.Thomas Girault, C.Cyrielle Mallart and D. H.Duc Hau Nguyen. Detecting Fake News Conspiracies with Multitask and Prompt-Based Learning.MediaEval 2021 - MediaEval Multimedia Evaluation benchmark. WorkshopOnline, NetherlandsDecember 2021, 1-3
  HALback to text
• 18 inproceedingsD.Deniz Engin, F.François Schnitzler, N. Q.Ngoc Q K Duong and Y.Yannis Avrithis. On the hidden treasure of dialog in video question answering.ICCV 2021 - IEEE/CVF International Conference on Computer VisionVirtual, FranceOctober 2021
  HALback to text
• 19 inproceedingsS. K.Suresh Kirthi Kumaraswamy, M.Miaojing Shi and E.Ewa Kijak. Detecting Human-Object Interaction with Mixed Supervision.WACV 2021 - Winter Conference on Applications of Computer VisionWaikoloa / Virtual, United StatesJanuary 2021, 1-10
  HALback to text
• 20 inproceedingsM.Michalis Lazarou, T.Tania Stathaki and Y.Yannis Avrithis. Iterative label cleaning for transductive and semi-supervised few-shot learning.ICCV 2021 - IEEE/CVF International Conference on Computer VisionVirtual, FranceOctober 2021
  HALback to text
• 21 inproceedingsG.Guillaume Le Noé Bienvenu, F.François Vignale, G.Guillaume Gravier and P.Pascale Sébillot. Utilisation d'approches automatiques pour la reconnaissance des expériences de lecture.Humanistica 2021 - Colloque de l'Association francophone des humanités numériquesProceedings of Humanistica 202Rennes, FranceMay 2021, 81-83
  HALback to text
• 22 inproceedingsC.Colin Leverger, T.Thomas Guyet, S.Simon Malinowski, V.Vincent Lemaire, A.Alexis Bondu, L.Laurence Rozé, A.Alexandre Termier and R.Régis Marguerie. Probabilistic forecasting of seasonal time series Combining clustering and classification for forecasting.ITISE 2021 - 7th International Conference on Time Series and ForecastingGran Canaria, SpainJuly 2021, 1-13
  HALback to text
• 23 inproceedingsY.Yann Lifchitz, Y.Yannis Avrithis and S.Sylvaine Picard. Few-Shot Few-Shot Learning and the role of Spatial Attention.ICPR 2020 - 25th International Conference on Pattern RecognitionVirtual, ItalyJanuary 2021, 1-7
  HALback to text
• 24 inproceedingsY.Yann Lifchitz, Y.Yannis Avrithis and S.Sylvaine Picard. Local Propagation for Few-Shot Learning.ICPR 2020 - 25th International Conference on Pattern RecognitionVirtual, ItalyJanuary 2021
  HALback to text
• 25 inproceedingsT.Thibault Maho, B.Benoît Bonnet, T.Teddy Furon and E.Erwan Le Merrer. RoBIC: A benchmark suite for assessing classifiers robustness.ICIP 2021 - IEEE International Conference on Image ProcessingAnchorage, Alaska, United StatesIEEESeptember 2021, 1-5
  HALDOIback to text
• 26 inproceedingsT.Thibault Maho, T.Teddy Furon and E. L.Erwan Le Merrer. SurFree: a fast surrogate-free black-box attack.CVPR 2021 - Conference on Computer Vision and Pattern RecognitionProc. of {IEEE} Conference on Computer Vision and Pattern Recognition, {CVPR}Virtual, FranceJune 2021, 10430--10439
  HALback to text
• 27 inproceedingsC.Cyrielle Mallart, M.Michel Le Nouy, G.Guillaume Gravier and P.Pascale Sébillot. Active Learning for Interactive Relation Extraction in a French Newspaper's Articles.RANLP 2021 - Recent Advances in Natural Language ProcessingProceedings of International Conference Recent Advances in Natural Language Processing RANLP 2021Online, BulgariaSeptember 2021, 886-894
  HALback to text
• 28 inproceedingsD. H.Duc Hau Nguyen, G.Guillaume Gravier and P.Pascale Sébillot. A Study of the Plausibility of Attention between RNN Encoders in Natural Language Inference.ICMLA 2021 - 20th IEEE International Conference on Machine Learning and ApplicationsPasadena, United StatesIEEEDecember 2021, 1-7
  HALback to text
• 29 inproceedingsO.Oriane Siméoni, M.Mateusz Budnik, Y.Yannis Avrithis and G.Guillaume Gravier. Rethinking deep active learning: Using unlabeled data at model training.International Conference on Pattern RecognitionMilan, ItalyJanuary 2021
  HALback to text
• 30 inproceedingsK.Karim Tit, T.Teddy Furon and M.Mathias Rousset. Efficient Statistical Assessment of Neural Network Corruption Robustness.NeurIPS 2021 - 35th Conference on Neural Information Processing Systems34Advances in Neural Information Processing Systems proceedingsSydney (virtual), AustraliaDecember 2021
  HALback to text
• 31 inproceedingsK.Karim Tit, T.Teddy Furon, M.Mathias Rousset and L.-M.Louis-Marie Traonouez. Évaluation statistique efficace de la robustesse de classifieurs.CAID 2021 - Conference on Artificial Intelligence for DefenseActes de la conférence CAID 2021Rennes, FranceNovember 2021, 1-11
  HALback to text
• 32 inproceedingsF.François Torregrossa, R.Robin Allesiardo, V.Vincent Claveau and G.Guillaume Gravier. Unsupervised Tree Extraction in Embedding Spaces for Taxonomy Induction.WI-IAT 2021 - 20th IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent TechnologyProceedings of the WI-IAT ConferenceMelbourne, AustraliaIEEEDecember 2021, 1-8
  HALDOIback to textback to text
• 33 inproceedingsF.François Vignale, G.Guillaume Le Noé Bienvenu, G.Guillaume Gravier and P.Pascale Sébillot. " Je pense que ça traite d'expérience de lecture,à voir ... " : retour sur une expérience d'annotation collaborative.Humanistica 2021 - Colloque annuel de l'association francophone des humanités numériquesProceedings of Humanistica 2021Rennes, FranceMay 2021, 84-85
  HALback to text
• 34 inproceedingsH.Hanwei Zhang, Y.Yannis Avrithis, T.Teddy Furon and L.Laurent Amsaleg. Patch Replacement: A Transformation-based Method to Improve Robustness against Adversarial Attacks.ACM Multimedia, Trustworthy AI WorkshopTrustworthy AI 2021 - 1st International Workshop on Trustworthy AI for Multimedia ComputingVirtual, ChinaACMOctober 2021, 1-10
  HALDOIback to text

Conferences without proceedings

• 35 inproceedingsA.Antoine Chaffin, V.Vincent Claveau and E.Ewa Kijak. PPL-MCTS: Constrained Textual Generation Through Discriminator-Guided Decoding.CtrlGen 2021 - Workshop on Controllable Generative Modeling in Language and Vision at NeurIPS 2021virtual, United StatesDecember 2021, 1-19
  HALback to textback to text

Scientific book chapters

• 36 inbookT.Teddy Furon. Traçage de traîtres.1Sécurité multimédia. Authentification et insertion de données cachéesISTEJune 2021, 201-228
  HALback to text
• 37 inbookH.Hanwei Zhang, T.Teddy Furon, L.Laurent Amsaleg and Y.Yannis Avrithis. Attaques et défenses de réseaux de neurones profonds : le cas de la classification d'images.Sécurité multimédia 1. Authentification et insertion de données cachéesJune 2021
  HALback to text

Edition (books, proceedings, special issue of a journal)

• 38 proceedingsF.François De VieillevilleS.Stéphane MayA.Adrien LagrangeA.A DupuisR.Rosa RuilobaF.Fred Ngolè MboulaT.Tristan Bitard-FeildelE.Erwan NoguesC.Corentin LarrocheJ.Johan MazelS.Stephan ClémençonR.Romain BurgotA.Alric GaurierL.Louis HulotL.Léo Isaac-DogninL.Laetitia LeichtnamE.Eric TotelN.Nicolas PrigentL.Ludovic MéR.Rémi BernhardP.-A.Pierre-Alain MoëllicJ.-M.Jean-Max DUTERTREK.Katarzyna KapustaV.Vincent ThouvenotO.Olivier BettanT.Tristan CharrierL.Luc BonnafouxF.-P.Francisco-Pierre PuigQ.Quentin LhoestT.Thomas RenaultA.Adrien BenamiraB.Benoit BonnetT.Teddy FuronP.Patrick BasB.Benjamin FarcyS.Silvia Gil-CasalsJ.Juliette MattioliM.Marc FiammanteM.Marc LambertR.Roman BressonJ.Johanne CohenE.Eyke HullermeierC.Christophe LabreucheM.Michele SebagT.Thomas ThebaudA.Anthony LarcherG.Gaël Le LanN.Nouredine NourR.Reda Belhaj-SoullamiC. L.Cédric L.R. BuronA.Alain PeresF.Frédéric BarbarescoA.Antoine d’AcremontG.Guillaume QuinA.Alexandre BaussardR.Ronan FabletL.Louis Morge-RolletF.Frederic Le RoyD.Denis Le JeuneR.Roland GautierB.Benjamin CamusE.Eric MonteuxM.Mikaël VermetA.Alex GoupilleauT.Tugdual CeillierM.-C.Marie-Caroline CorbineauProceedings of the Conference on Artificial Intelligence for Defence 2020.CAID 2020 - Second Conference on Artificial Intelligence for DefenceRennes, FranceApril 2021
  HAL

Doctoral dissertations and habilitation theses

• 39 thesisC. B.Cheikh Brahim El Vaigh. Content and data linking leveraging ontological knowledge in data journalism.Université Rennes 1January 2021
  HALback to text
• 40 thesisM.Marzieh Gheisari. Secure identification for the Internet of Things.Inria Rennes - Bretagne AtlantiqueJuly 2021
  HALback to text
• 41 thesisY.Yann Lifchitz. Making the most of available data : representation and adaptation for few-shot image classification.Université Rennes 1April 2021
  HALback to text
• 42 thesisH.Hanwei Zhang. Deep Learning in Adversarial Context.École normale supérieure de RennesJune 2021
  HALback to text

Reports & preprints

• 43 miscA.Antoine Chaffin, V.Vincent Claveau and E.Ewa Kijak. Generating texts under constraint through discriminator-guided MCTS.November 2021
  HALback to textback to text
• 44 miscV.Vincent Claveau, A.Antoine Chaffin and E.Ewa Kijak. Generating artificial texts as substitution or complement of training data.November 2021
  HALback to text
• 45 miscV.Vincent Claveau. Query expansion with artificially generated texts.November 2021
  HALback to text
• 46 miscT.Teddy Furon. Note: An alternative proof of the vulnerability of $k$-NN classifiers in high intrinsic dimensionality regions.January 2021
  HAL

Other scientific publications

• 47 articleG.Guillaume Gravier, E.Elisa Fromont, N.Nicolas Courty, T.Teddy Furon, C.Christine Guillemot and P.Paolo Robuffo Giordano. Rennes - une IA souveraine au service de la vie publique.Bulletin de l'Association Française pour l'Intelligence Artificielle2021
  HALback to text

Scientific popularization

• 48 inproceedingsS.Shashanka Venkataramanan, B.Bill Psomas, E.Ewa Kijak, L.Laurent Amsaleg, K.Konstantinos Karantzalos and Y.Yannis Avrithis. It Takes Two to Tango: Mixup for Deep Metric Learning.ICLR 2022 - 10th International Conference on Learning RepresentationsVirtual, FranceApril 2022, 1-21
  HAL