The main result of this paper is to prove the existence of a finite basis in the description logi... more The main result of this paper is to prove the existence of a finite basis in the description logic ALC. We show that the set of General Concept Inclusions (GCIs) holding in a finite model has always a finite basis, i.e. these GCIs can be derived from finitely many of the GCIs. This result extends a previous result from Baader and Distel, which showed the existence of a finite basis for GCIs holding in a finite model but for the inexpressive description logics E L and E L gf p. We also provide an algorithm for computing this finite basis, and prove its correctness. As a byproduct, we extend our finite basis theorem to any finitely generated complete covariety (i.e. any class of models closed under morphism domain, coproduct and quotient, and generated from a finite set of finite models).
Unsupervised Domain Adaptation aims to learn a model on a source domain with labeled data in orde... more Unsupervised Domain Adaptation aims to learn a model on a source domain with labeled data in order to perform well on unlabeled data of a target domain. Current approaches focus on learning Domain Invariant Representations. It relies on the assumption that such representations are wellsuited for learning the supervised task in the target domain. We rather believe that a better and minimal assumption for performing Domain Adaptation is the Hidden Covariate Shift hypothesis. Such approach consists in learning a representation of the data such that the label distribution conditioned on this representation is domain invariant. From the Hidden Covariate Shift assumption, we derive an optimization procedure which learns to match an estimated joint distribution on the target domain and a re-weighted joint distribution on the source domain. The re-weighting is done in the representation space and is learned during the optimization procedure. We show on synthetic data and real world data that our approach deals with both Target Shift and Concept Drift. We report state-of-the-art performances on Amazon Reviews dataset (Blitzer et al., 2007) demonstrating the viability of this approach.
When we can not assume a large amount of annotated data , active learning is a good strategy. It ... more When we can not assume a large amount of annotated data , active learning is a good strategy. It consists in learning a model on a small amount of annotated data (annotation budget) and in choosing the best set of points to annotate in order to improve the previous model and gain in generalization. In deep learning, active learning is usually implemented as an iterative process in which successive deep models are updated via fine tuning, but it still poses some issues. First, the initial batch of annotated images has to be sufficiently large to train a deep model. Such an assumption is strong, especially when the total annotation budget is reduced. We tackle this issue by using an approach inspired by transfer learning. A pre-trained model is used as a feature extractor and only shallow classifiers are learned during the active iterations. The second issue is the effectiveness of probability or feature estimates of early models for AL task. Samples are generally selected for annotat...
The main result of this paper is to prove the existence of a finite basis in the description logi... more The main result of this paper is to prove the existence of a finite basis in the description logic ALC. We show that the set of General Concept Inclusions (GCIs) holding in a finite model has always a finite basis, i.e. these GCIs can be derived from finitely many of the GCIs. This result extends a previous result from Baader and Distel, which showed the existence of a finite basis for GCIs holding in a finite model but for the inexpressive description logics EL and EL_gfp. We also provide an algorithm for computing this finite basis, and prove its correctness. As a byproduct, we extend our finite basis theorem to any finitely generated complete covariety (i.e. any class of models closed under morphism domain, coproduct and quotient, and generated from a finite set of finite models).
Many real-world visual recognition use-cases can not directly benefit from state-of-the-art CNN-b... more Many real-world visual recognition use-cases can not directly benefit from state-of-the-art CNN-based approaches because of the lack of many annotated data. The usual approach to deal with this is to transfer a representation pre-learned on a large annotated source-task onto a target-task of interest. This raises the question of how well the original representation is "universal", that is to say directly adapted to many different target-tasks. To improve such universality, the state-of-the-art consists in training networks on a diversified source problem, that is modified either by adding generic or specific categories to the initial set of categories. In this vein, we proposed a method that exploits finer-classes than the most specific ones existing, for which no annotation is available. We rely on unsupervised learning and a bottom-up split and merge strategy. We show that our method learns more universal representations than state-of-the-art, leading to significantly be...
With the recent successes of black-box models in Artificial Intelligence (AI) and the growing int... more With the recent successes of black-box models in Artificial Intelligence (AI) and the growing interactions between humans and AIs, explainability issues have risen. In this article, in the context of high-stake applications, we propose an approach for explainable classification and annotation of images. It is based on a transparent model, whose reasoning is accessible and human understandable, and on interpretable fuzzy relations that enable to express the vagueness of natural language. The knowledge about relations is set beforehand by an expert and thus training instances do not need to be annotated. The most relevant relations are extracted using a fuzzy frequent itemset mining algorithm in order to build rules, for classification, and constraints, for annotation. We also present two heuristics that make the process of evaluating relations faster. Since the strengths of our approach are the transparency of the model and the interpretability of the relations, an explanation in natural language can be generated. Supported by experimental results, we show that, given a segmentation of the input, our approach is able to successfully perform the target task and generate explanations that were judged as consistent and convincing by a set of participants.
We propose an original way of enriching Description Logics with ab- duction reasoning services by... more We propose an original way of enriching Description Logics with ab- duction reasoning services by computing the best explanations of an observation through mathematical morphology (using erosions) over the Concept Lattice of a background theory. The intended application is scene understanding and spatial
Dans cet article, nous proposons d’etendre les liens deja etablis entre les operateurs de derivat... more Dans cet article, nous proposons d’etendre les liens deja etablis entre les operateurs de derivation utilises en analyse formelle de concepts et des operateurs de morphologie mathematique a l’analyse formelle de concepts flous. De plus, nous proposons d’exploiter la morphologie mathematique pour naviguer dans le treillis de concepts flous et raisonner sur ceux-ci. Cet article propose a la fois une discussion generale ainsi que de nouveaux resultats sur ces liens et leur interet potentiel.
Resume : L’objet de cet article est d’enrichir les logiques de description par des services de ra... more Resume : L’objet de cet article est d’enrichir les logiques de description par des services de raisonnement abductif a des fin d’interpretation d’images. Dans un cadre algebrique, nous mettons en synergie des ingredients provenant de la morphologie mathematique, des logiques de description et de l’analyse formelle de concepts. Plus precisement, nous proposons de calculer la « meilleure » explication d’une observation donnee a partir d’une succession d’erosions algebriques sur le treillis de concepts associe a la theorie du domaine. Nous montrons que les operateurs ainsi definis satisfont les postulats de rationalite du raisonnement abductif. Comme illustration, nous considerons le domaine de l’interpretation d’images a partir de modeles, ou des connaissances a priori structurees beneficient de representations ontologiques et des logiques de description. Nous formulons donc le probleme d’interpretation a base de modele comme un probleme d’abduction : l’image represente l’observation ...
Recent deep generative models are able to provide photo-realistic images as well as visual or tex... more Recent deep generative models are able to provide photo-realistic images as well as visual or textual content embeddings useful to address various tasks of computer vision and natural language processing. Their usefulness is nevertheless often limited by the lack of control over the generative process or the poor understanding of the learned representation. To overcome these major issues, very recent work has shown the interest of studying the semantics of the latent space of generative models. In this paper, we propose to advance on the interpretability of the latent space of generative models by introducing a new method to find meaningful directions in the latent space of any generative model along which we can move to control precisely specific properties of the generated image like the position or scale of the object in the image. Our method does not require human annotations and is particularly well suited for the search of directions encoding simple transformations of the gene...
Despite the recent successes of deep learning, such models are still far from some human abilitie... more Despite the recent successes of deep learning, such models are still far from some human abilities like learning from few examples, reasoning and explaining decisions. In this paper, we focus on organ annotation in medical images and we introduce a reasoning framework that is based on learning fuzzy relations on a small dataset for generating explanations. Given a catalogue of relations, it efficiently induces the most relevant relations and combines them for building constraints in order to both solve the organ annotation task and generate explanations. We test our approach on a publicly available dataset of medical images where several organs are already segmented. A demonstration of our model is proposed with an example of explained annotations. It was trained on a small training set containing as few as a couple of examples.
Deep learning approaches are successful in a wide range of AI problems and in particular for visu... more Deep learning approaches are successful in a wide range of AI problems and in particular for visual recognition tasks. However, there are still open problems among which is the capacity to handle streams of visual information and the management of class imbalance in datasets. Existing research approaches these two problems separately while they co-occur in real world applications. Here, we study the problem of learning incrementally from imbalanced datasets. We focus on algorithms which have a constant deep model complexity and use a bounded memory to store exemplars of old classes across incremental states. Since memory is bounded, old classes are learned with fewer images than new classes and an imbalance due to incremental learning is added to the initial dataset imbalance. A score prediction bias in favor of new classes appears and we evaluate a comprehensive set of score calibration methods to reduce it. Evaluation is carried with three datasets, using two dataset imbalance con...
Many real-world visual recognition use-cases can not directly benefit from state-of-the-art CNN-b... more Many real-world visual recognition use-cases can not directly benefit from state-of-the-art CNN-based approaches because of the lack of many annotated data. The usual approach to deal with this is to transfer a representation pre-learned on a large annotated source-task onto a target-task of interest. This raises the question of how well the original representation is "universal", that is to say directly adapted to many different target-tasks. To improve such universality, the state-of-the-art consists in training networks on a diversified source problem, that is modified either by adding generic or specific categories to the initial set of categories. In this vein, we proposed a method that exploits finer-classes than the most specific ones existing, for which no annotation is available. We rely on unsupervised learning and a bottom-up split and merge strategy. We show that our method learns more universal representations than state-of-the-art, leading to significantly be...
2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2018
The goal of explainable artificial intelligence is to solve problems in a way that humans can und... more The goal of explainable artificial intelligence is to solve problems in a way that humans can understand how it does it. However, few approaches have been proposed so far and some of them lay more emphasis on interpretability than on explainability. In this paper, we propose an approach that is based on learning fuzzy relations and fuzzy properties. We extract frequent relations from a dataset to generate an explained decision. Our approach can deal with different problems, such as classification or annotation. A model was built to perform explained classification on a toy dataset that we generated. It managed to correctly classify examples while providing convincing explanations. A few areas for improvement have been spotted, such as the need to filter relations and properties before or while learning them in order to avoid useless computations.
International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 2016
In this paper we extend some previously established links between the derivation operators used i... more In this paper we extend some previously established links between the derivation operators used in formal concept analysis and some mathematical morphology operators to fuzzy concept analysis. We also propose to use mathematical morphology to navigate in a fuzzy concept lattice and perform operations on it. Links with other lattice-based for malisms such as rough sets and F-transforms are also established. This paper proposes a discussion and new results on such links and their potential interest.
Ce papier 1 traite du problème de la construction d'une ontologie commune pour un ensemble d'onto... more Ce papier 1 traite du problème de la construction d'une ontologie commune pour un ensemble d'ontologies de domaine afin de permettre leur partage et leur intégration dans une plate-forme collaborative. Nous proposons un nouvel algorithme hiérarchique pour représenter les concepts comme des ensembles flous à l'aide d'une ontologie de référence. Contrairement aux représentations originales des concepts basées sur des instances, cela permet l'application de méthodes de raisonnement flou dans le but de caractériser et de mesurer le degré des relations entre les concepts des ontologies de domaine. Nous proposons une application de l'approche dans le domaine du multimédia.
Integrating Bipolar Fuzzy Mathematical Morphology in Description Logics for Spatial Reasoning
Bipolarity is an important feature of spatial information, involved in the expression of preferen... more Bipolarity is an important feature of spatial information, involved in the expression of preferences and constraints about spatial positioning or in pairs of opposite spatial relations such as left and right. Another important feature is imprecision which has to be taken into account to model vagueness, inherent to many spatial relations (as for instance vague expressions such as close to, to the right of), and to gain in robustness in the representations. In previous works, we have shown that fuzzy sets and fuzzy mathematical morphology are appropriate frameworks, on the one hand, to represent bipolarity and imprecision of spatial relations and, on the other hand, to combine qualitative and quantitative reasoning in description logics extended with fuzzy concrete domains. The purpose of this paper is to integrate the bipolarity feature in the latter logical framework based on bipolar and fuzzy mathematical morphology and description logics with fuzzy concrete domains. Two important...
In this paper we highlight a few features of the semantic gap problem in image interpretation. We... more In this paper we highlight a few features of the semantic gap problem in image interpretation. We show that semantic image interpretation can be seen as a symbol grounding problem. In this context, ontologies provide a powerful framework to represent domain knowledge, concepts and their relations, and to reason about them. They are likely to be more and more developed for image interpretation. A lot of image interpretation systems rely strongly on descriptions of objects through their characteristics such as shape, location, image intensities. However, spatial relations are very important too and provide a structural description of the imaged phenomenon, which is often more stable and less prone to variability than pure object descriptions. We show that spatial relations can be integrated in domain ontologies. Because of the intrinsic vagueness we have to cope with, at different levels (image objects, spatial relations, variability, questions to be answered, etc.), fuzzy representations are well adapted and provide a consistent formal framework to address this key issue, as well as the associated reasoning and decision making aspects. Our view is that ontology-based methods can be very useful for image interpretation if they are associated to operational models relating the ontology concepts to image information. In particular, we propose operational models of spatial relations, based on fuzzy representations.
In image interpretation and computer vision, spatial relations between objects and spatial reason... more In image interpretation and computer vision, spatial relations between objects and spatial reasoning are of prime importance for recognition and interpretation tasks. Quantitative representations of spatial knowledge have been proposed in the literature. In the Artificial Intelligence community, logical formalisms such as ontologies have also been proposed for spatial knowledge representation and reasoning, and a challenging and open problem consists in bridging the gap between these ontological representations and the quantitative ones used in image interpretation. In this paper, we propose a new description logic, named ALC(F), dedicated to spatial reasoning for image understanding. Our logic relies on the family of description logics equipped with concrete domains, a widely accepted way to integrate quantitative and qualitative qualities of real world objects in the conceptual domain, in which we have integrated mathematical morphological operators as predicates. Merging description logic with mathematical morphology enables us to provide new mechanisms to derive useful concrete representations of spatial concepts and new qualitative and quantitative spatial reasoning tools. It also enables imprecision and uncertainty of spatial knowledge to be taken into account through the fuzzy representation of spatial relations. We illustrate the benefits of our formalism on a model-guided cerebral image interpretation task.
Proceedings of the 3rd ACM conference on International conference on multimedia retrieval, 2013
We address the problem of tag completion for automatic image annotation. Our method consists in t... more We address the problem of tag completion for automatic image annotation. Our method consists in two main steps: creating a list of "candidate tags" from the visual neighbors of the untagged image then using them as pieces of evidence to be combined to provide the final list of predicted tags. Both steps introduce a scheme to tackle with imprecision and uncertainty. First, a bag-of-words (BOW) signature is generated for each neighbor using local soft coding. Second, a sum-pooling operation across the BOW of the k nearest neighbors provides the list of "candidate tags". Finally, we use neighbors as pieces of evidence to be combined according to the Dempster's rule to predict the more relevant tags. The method is evaluated in the context of image classification and that of tag suggestion. The database used for visual neighbors search contains 1.2 million images extracted from Flickr. Classification is evaluated on the well known Pascal VOC 2007 and MIR Flickr datasets, on which we obtain similar or better results than the state-of-the-art. For tag suggestion, we manually annotated 241 queries. As well, we obtain competitive results on this task.
Uploads
Papers by Celine Hudelot