Centre Inria de l'Université de Lorraine

OBJECTIVE: Most current electroencephalography (EEG)-based brain-computer interfaces (BCIs) are based on machine learning algorithms. There is a large diversity of classifier types that are used in this field, as described in our 2007 review paper. Now, approximately ten years after this review publication, many new algorithms have been developed and tested to classify EEG signals in BCIs. The time is therefore ripe for an updated review of EEG classification algorithms for BCIs. APPROACH: We surveyed the BCI and machine learning literature from 2007 to 2017 to identify the new classification approaches that have been investigated to design BCIs. We synthesize these studies in order to present such algorithms, to report how they were used for BCIs, what were the outcomes, and to identify their pros and cons. MAIN RESULTS: We found that the recently designed classification algorithms for EEG-based BCIs can be divided into four main categories: adaptive classifiers, matrix and tensor classifiers, transfer learning and deep learning, plus a few other miscellaneous classifiers. Among these, adaptive classifiers were demonstrated to be generally superior to static ones, even with unsupervised adaptation. Transfer learning can also prove useful although the benefits of transfer learning remain unpredictable. Riemannian geometry-based methods have reached state-of-the-art performances on multiple BCI problems and deserve to be explored more thoroughly, along with tensor-based methods. Shrinkage linear discriminant analysis and random forests also appear particularly useful for small training samples settings. On the other hand, deep learning methods have not yet shown convincing improvement over state-of-the-art BCI methods. SIGNIFICANCE: This paper provides a comprehensive overview of the modern classification algorithms used in EEG-based BCIs, presents the principles of these methods and guidelines on when and how to use them. It also identifies a number of challenges to further advance EEG classification in BCI.

The topic of this review is geometric registration in robotics. Registration algorithms associate sets of data into a common coordinate system. They have been used extensively in object reconstruction, inspection, medical application, and localization of mobile robotics. We focus on mobile robotics applications in which point clouds are to be registered. While the underlying principle of those algorithms is simple, many variations have been proposed for many different applications. In this review, we give a historical perspective of the registration problem and show that the plethora of solutions can be organized and differentiated according to a few elements. Accordingly, we present a formalization of geometric registration and cast algorithms proposed in the literature into this framework. Finally, we review a few applications of this framework in mobile robotics that cover different kinds of platforms, environments, and tasks. These examples allow us to study the specific requirements of each use case and the necessary configuration choices leading to the registration implementation. Ultimately, the objective of this review is to provide guidelines for the choice of geometric registration configuration.

Speech enhancement and separation are core problems in audio signal processing, with commercial applications in devices as diverse as mobile phones, conference call systems, hands-free systems, or hearing aids. In addition, they are crucial preprocessing steps for noise-robust automatic speech and speaker recognition. Many devices now have two to eight microphones. The enhancement and separation capabilities offered by these multichannel interfaces are usually greater than those of single-channel interfaces. Research in speech enhancement and separation has followed two convergent paths, starting with microphone array processing and blind source separation, respectively. These communities are now strongly interrelated and routinely borrow ideas from each other. Yet, a comprehensive overview of the common foundations and the differences between these approaches is lacking at present. In this paper, we propose to fill this gap by analyzing a large number of established and recent techniques according to four transverse axes: 1) the acoustic impulse response model, 2) the spatial filter design criterion, 3) the parameter estimation algorithm, and 4) optional postfiltering. We conclude this overview paper by providing a list of software and data resources and by discussing perspectives and future trends in the field.

BACKGROUND: The Critical Assessment of Functional Annotation (CAFA) is an ongoing, global, community-driven effort to evaluate and improve the computational annotation of protein function. RESULTS: Here, we report on the results of the third CAFA challenge, CAFA3, that featured an expanded analysis over the previous CAFA rounds, both in terms of volume of data analyzed and the types of analysis performed. In a novel and major new development, computational predictions and assessment goals drove some of the experimental assays, resulting in new functional annotations for more than 1000 genes. Specifically, we performed experimental whole-genome mutation screening in Candida albicans and Pseudomonas aureginosa genomes, which provided us with genome-wide experimental data for genes associated with biofilm formation and motility. We further performed targeted assays on selected genes in Drosophila melanogaster, which we suspected of being involved in long-term memory. CONCLUSION: We conclude that while predictions of the molecular function and biological process annotations have slightly improved over time, those of the cellular component have not. Term-centric prediction of experimental annotations remains equally challenging; although the performance of the top methods is significantly better than the expectations set by baseline methods in C. albicans and D. melanogaster, it leaves considerable room and need for improvement. Finally, we report that the CAFA community now involves a broad range of participants with expertise in bioinformatics, biological experimentation, biocuration, and bio-ontologies, working together to improve functional annotation, computational function prediction, and our ability to manage big data in the era of large experimental screens.

One of the challenges in geometry processing is to automatically reconstruct a higher-level representation from raw geometric data. For instance, computing a parameterization of an object helps attaching information to it and converting between various representations. More generally, this family of problems may be thought of in terms of constructing structured function bases attached to surfaces. In this paper, we study a specific type of hierarchical function bases, defined by the eigenfunctions of the Laplace-Beltrami operator. When applied to a sphere, this function basis corresponds to the classical spherical harmonics. On more general objects, this defines a function basis well adapted to the geometry and the topology of the object. Based on physical analogies (vibration modes), we first give an intuitive view before explaining the underlying theory. We then explain in practice how to compute an approximation of the eigenfunctions of a differential operator, and show possible applications in geometry processing

YASS is a DNA local alignment tool based on an efficient and sensitive filtering algorithm. It applies transition-constrained seeds to specify the most probable conserved motifs between homologous sequences, combined with a flexible hit criterion used to identify groups of seeds that are likely to exhibit significant alignments. A web interface (http://www.loria.fr/projects/YASS/) is available to upload input sequences in fasta format, query the program and visualize the results obtained in several forms (dot-plot, tabular output and others). A standalone version is available for download from the web page.

MOTIVATION: Modelling protein-protein interactions (PPIs) is an increasingly important aspect of structural bioinformatics. However, predicting PPIs using in silico docking techniques is computationally very expensive. Developing very fast protein docking tools will be useful for studying large-scale PPI networks, and could contribute to the rational design of new drugs. RESULTS: The Hex spherical polar Fourier protein docking algorithm has been implemented on Nvidia graphics processor units (GPUs). On a GTX 285 GPU, an exhaustive and densely sampled 6D docking search can be calculated in just 15 s using multiple 1D fast Fourier transforms (FFTs). This represents a 45-fold speed-up over the corresponding calculation on a single CPU, being at least two orders of magnitude times faster than a similar CPU calculation using ZDOCK 3.0.1, and estimated to be at least three orders of magnitude faster than the GPU-accelerated version of PIPER on comparable hardware. Hence, for the first time, exhaustive FFT-based protein docking calculations may now be performed in a matter of seconds on a contemporary GPU. Three-dimensional Hex FFT correlations are also accelerated by the GPU, but the speed-up factor of only 2.5 is much less than that obtained with 1D FFTs. Thus, the Hex algorithm appears to be especially well suited to exploit GPUs compared to conventional 3D FFT docking approaches. AVAILABILITY: http://hex.loria.fr/ and http://hexserver.loria.fr/.

Horizontal transfer of genomic islands (GEIs), that is, chromosomal regions encoding functions that can be advantageous for the host, plays a key role in bacterial evolution, but their mechanisms of transfer remained elusive for a long time. Recent data suggest that numerous GEIs belong to noncanonical classes of mobile genetic elements (MGEs) that can transfer by conjugation. Among them, the integrative and conjugative elements encode their own excision, conjugative transfer, and integration, whereas the integrative mobilizable elements are autonomous for excision and integration but require the conjugation machinery of helper elements to transfer. Others can self-transfer but require the recombination machinery of the recipient cell to integrate. All these MGEs evolve by acquisition, deletion, or exchange of modules, that is, groups of genes involved in the same function. Moreover, composite GEIs can result from the insertion of a MGE within another or from the site-specific integration of an incoming MGE into one of the recombination sites flanking a resident GEI (tandem accretion). Tandem accretion enables the cis-conjugative mobilization of highly degenerated and nonautonomous GEIs, the cis-mobilizable elements. All these mechanisms contribute to the plasticity and complex evolution of GEIs and explain the highly diverse tableau revealed by more and more genome comparisons.

Multi-microphone arrays allow for the use of spatial filtering techniques that can greatly improve noise reduction and source separation. However, for speech and audio data, work on noise reduction or separation has focused primarily on one- or two-channel systems. Because of this, databases of multichannel environmental noise are not widely available. DEMAND (Diverse Environments Multi-channel Acoustic Noise Database) addresses this problem by providing a set of 16-channel noise files recorded in a variety of indoor and outdoor settings. The data was recorded using a planar microphone array consisting of four staggered rows, with the smallest distance between microphones being 5 cm and the largest being 21.8 cm. DEMAND is freely available under a Creative Commons license to encourage research into algorithms beyond the stereo setup.

Centroidal Voronoi tessellation (CVT) is a particular type of Voronoi tessellation that has many applications in computational sciences and engineering, including computer graphics. The prevailing method for computing CVT is Lloyd's method, which has linear convergence and is inefficient in practice. We develop new efficient methods for CVT computation and demonstrate the fast convergence of these methods. Specifically, we show that the CVT energy function has 2nd order smoothness for convex domains with smooth density, as well as in most situations encountered in optimization. Due to the 2nd order smoothness, it is possible to minimize the CVT energy functions using Newton-like optimization methods and expect fast convergence. We propose a quasi-Newton method to compute CVT and demonstrate its faster convergence than Lloyd's method with various numerical examples. It is also significantly faster and more robust than the Lloyd-Newton method, a previous attempt to accelerate CVT. We also demonstrate surface remeshing as a possible application.

Distant-microphone automatic speech recognition (ASR) remains a challenging goal in everyday environments involving multiple background sources and reverberation. This paper is intended to be a reference on the 2nd `CHiME' Challenge, an initiative designed to analyze and evaluate the performance of ASR systems in a real-world domestic environment. Two separate tracks have been proposed: a small-vocabulary task with small speaker movements and a medium-vocabulary task without speaker movements. We discuss the rationale for the challenge and provide a detailed description of the datasets, tasks and baseline performance results for each track.

A repetition in a word w is a subword with the period of at most half of the subword length. We study maximal repetitions occurring in w, that is those for which any extended subword of w has a bigger period. The set of such repetitions represents in a compact way all repetitions in w. We first prove a combinatorial result asserting that the sum of exponents of all maximal repetitions of a word of length n is bounded by a linear function in n. This implies, in particular that there is only a linear number of maximal repetitions in a word. This allows us to construct a linear-time algorithm for finding all maximal repetitions. Some consequences and applications of these results are discussed, as well as related works.

This article addresses the problem of multichannel audio source separation. We propose a framework where deep neural networks (DNNs) are used to model the source spectra and combined with the classical multichannel Gaussian model to exploit the spatial information. The parameters are estimated in an iterative expectation-maximization (EM) fashion and used to derive a multichannel Wiener filter. We present an extensive experimental study to show the impact of different design choices on the performance of the proposed technique. We consider different cost functions for the training of DNNs, namely the probabilistically motivated Itakura-Saito divergence, and also Kullback-Leibler, Cauchy, mean squared error, and phase-sensitive cost functions. We also study the number of EM iterations and the use of multiple DNNs, where each DNN aims to improve the spectra estimated by the preceding EM iteration. Finally, we present its application to a speech enhancement problem. The experimental results show the benefit of the proposed multichannel approach over a single-channel DNN-based approach and the conventional multichannel nonnegative matrix factorization-based iterative EM algorithm.

Intra-annual radial growth variations of two Norway spruce trees (Picea abies (L.) Karst.) were monitored over 4 years, at four heights up the stem, by means of point-dendrometers. The trees were then felled and radial wood samples were cut from the radii that had been monitored by the dendrometers and analyzed for density. From the radial growth measurements recorded by the dendrometers, we related positions within the rings to dates, thus making possible investigation of the relationships between changes within the rings in wood density and fluctuations in climate or growth rate. Radial growth started in early April and ended, with large intra-annual differences, in August or September. Short-term variations in growth rate were related to fluctuations in climate parameters and soil water reserves. The sensitivity of radial growth to climate decreased with stem height. Wood density responded strongly to drought events, and a dry period in June 1996 induced false-ring formation. Wood density was relatively independent of growth rate and climatic conditions during the first part of the growing season, but increased with decreasing radial growth rate later in the growing season.

Despite the growth of prevention techniques, phishing remains an important threat since the principal countermeasures in use are still based on reactive URL blacklisting. This technique is inefficient due to the short lifetime of phishing Web sites, making recent approaches relying on real-time or proactive phishing URL detection techniques more appropriate. In this paper, we introduce PhishStorm, an automated phishing detection system that can analyze in real time any URL in order to identify potential phishing sites. PhishStorm can interface with any email server or HTTP proxy. We argue that phishing URLs usually have few relationships between the part of the URL that must be registered (low-level domain) and the remaining part of the URL (upper-level domain, path, query). We show in this paper that experimental evidence supports this observation and can be used to detect phishing sites. For this purpose, we define the new concept of intra-URL relatedness and evaluate it using features extracted from words that compose a URL based on query data from Google and Yahoo search engines. These features are then used in machine-learning-based classification to detect phishing URLs from a real dataset. Our technique is assessed on 96 018 phishing and legitimate URLs that result in a correct classification rate of 94.91% with only 1.44% false positives. An extension for a URL phishingness rating system exhibiting high confidence rate ( $>$ 99%) is proposed. We discuss in this paper efficient implementation patterns that allow real-time analytics using Big Data architectures such as STORM and advanced data structures based on the Bloom filter.

"Electronic-sport" (E-Sport) is now established as a new entertainment genre. More and more players enjoy streaming their games, which attract even more viewers. In fact, in a recent social study, casual players were found to prefer watching professional gamers rather than playing the game themselves. Within this context, advertising provides a significant source of revenue to the professional players, the casters (displaying other people's games) and the game streaming platforms. For this paper, we crawled, during more than 100 days, the most popular among such specialized platforms: Twitch.tv. Thanks to these gigabytes of data, we propose a first characterization of a new Web community, and we show, among other results, that the number of viewers of a streaming session evolves in a predictable way, that audience peaks of a game are explainable and that a Condorcet method can be used to sensibly rank the streamers by popularity. Last but not least, we hope that this paper will bring to light the study of E-Sport and its growing community. They indeed deserve the attention of industrial partners (for the large amount of money involved) and researchers (for interesting problems in social network dynamics, personalized recommendation, sentiment analysis, etc.).

Content Centric Networking (CCN) has recently emerged as a promising architecture to deliver content at large-scale. It is based on named-data where a packet address names content and not its location. Then, the premise is to cache content on the network nodes along the delivery path. An important feature for CCN is therefore to manage the cache of the nodes. In this paper, we present Most Popular Content (MPC), a new caching strategy adapted to CCN networks. By caching only popular content, we show through extensive simulation experiments that MPC is able to cache less content while, at the same time, it still achieves a higher Cache Hit and outperforms existing default caching strategy in CCN.

Abstract Hybrid languages are expansions of propositional modal languages which can refer to (or even quantify over) worlds. The use of strong hybrid languages dates back to at least [Pri67], but recent work (for example [BS98, BT98a, BT99]) has focussed on a more constrained system called H (↓, @). We show in detail that (↓, @) is modally natural. We begin by studying its expressivity, and provide model theoretic characterizations (via a restricted notion of Ehrenfeucht-Fraïssé game, and an enriched notion of bisimulation) and a syntactic characterization (in terms of bounded formulas). The key result to emerge is that (↓, @) corresponds to the fragment of first-order logic which is invariant for generated submodels. We then show that (↓, @) enjoys (strong) interpolation, provide counterexamples for its finite variable fragments, and show that weak interpolation holds for the sublanguage (@). Finally, we provide complexity results for (@) and other fragments and variants, and sharpen known undecidability results for (↓, @).

Abstract We propose a new isotropic remeshing method, based on Centroidal Voronoi Tessellation (CVT) . Constructing CVT requires to repeatedly compute Restricted Voronoi Diagram (RVD) , defined as the intersection between a 3D Voronoi diagram and an input mesh surface. Existing methods use some approximations of RVD. In this paper, we introduce an efficient algorithm that computes RVD exactly and robustly. As a consequence, we achieve better remeshing quality than approximation‐based approaches, without sacrificing efficiency. Our method for RVD computation uses a simple procedure and a kd ‐tree to quickly identify and compute the intersection of each triangle face with its incident Voronoi cells. Its time complexity is O ( m log n ), where n is the number of seed points and m is the number of triangles of the input mesh. Fast convergence of CVT is achieved using a quasi‐Newton method, which proved much faster than Lloyd's iteration. Examples are presented to demonstrate the better quality of remeshing results with our method than with the state‐of‐art approaches.

Many algorithms in texture synthesis, nonphotorealistic rendering (hatching), or remeshing require to define the orientation of some features (texture, hatches, or edges) at each point of a surface. In early works, tangent vector (or tensor) fields were used to define the orientation of these features. Extrapolating and smoothing such fields is usually performed by minimizing an energy composed of a smoothness term and of a data fitting term. More recently, dedicated structures ( N -RoSy and N -symmetry direction fields ) were introduced in order to unify the manipulation of these fields, and provide control over the field's topology (singularities). On the one hand, controlling the topology makes it possible to have few singularities, even in the presence of high frequencies (fine details) in the surface geometry. On the other hand, the user has to explicitly specify all singularities, which can be a tedious task. It would be better to let them emerge naturally from the direction extrapolation and smoothing. This article introduces an intermediate representation that still allows the intuitive design operations such as smoothing and directional constraints, but restates the objective function in a way that avoids the singularities yielded by smaller geometric details. The resulting design tool is intuitive, simple, and allows to create fields with simple topology, even in the presence of high geometric frequencies. The generated field can be used to steer global parameterization methods (e.g., QuadCover).