The desirability of the mid-band frequency range (1 - 10 GHz) for federal and commercial applications, combined with the growing applications for commercial indoor use-cases, such as factory automation, opens up a new approach to spectrum sharing: the same frequency bands used outdoors by federal incumbents can be reused by commercial indoor users. A recent example of such sharing, between commercial systems, is the 6 GHz band (5.925 - 7.125 GHz) where unlicensed, low-power-indoor (LPI) users share the band with outdoor incumbents, primarily fixed microwave links. However, to date, there exist no reliable, automatic means of determining whether a device is indoors or outdoors, necessitating the use of other mechanisms such as mandating indoor access points (APs) to have integrated antennas and not be battery powered, and reducing transmit power of client devices which may be outdoors. An accurate indoor/outdoor (I/O) classification addresses these challenges, enabling automatic transmit power adjustments without interfering with incumbents. To this end, we leverage the Global Navigation Satellite System (GNSS) signals for I/O classification. GNSS signals, designed inherently for outdoor reception and highly susceptible to indoor attenuation and blocking, provide a robust and distinguishing feature for environmental sensing. We develop various methodologies, including threshold-based techniques and machine learning approaches and evaluate them using an expanded dataset gathered from diverse geographical locations. Our results demonstrate that GNSS-based methods alone can achieve greater accuracy than approaches relying solely on wireless (Wi-Fi) data, particularly in unfamiliar locations. Furthermore, the integration of GNSS data with Wi-Fi information leads to improved classification accuracy, showcasing the significant benefits of multi-modal data fusion.

Succinct representation of complex signals using coordinate-based neural representations (CNRs) has seen great progress, and several recent efforts focus on extending them for handling videos. Here, the main challenge is how to (a) alleviate a compute-inefficiency in training CNRs to (b) achieve high-quality video encoding while (c) maintaining the parameter-efficiency. To meet all requirements (a), (b), and (c) simultaneously, we propose neural video representations with learnable positional features (NVP), a novel CNR by introducing "learnable positional features" that effectively amortize a video as latent codes. Specifically, we first present a CNR architecture based on designing 2D latent keyframes to learn the common video contents across each spatio-temporal axis, which dramatically improves all of those three requirements. Then, we propose to utilize existing powerful image and video codecs as a compute-/memory-efficient compression procedure of latent codes. We demonstrate the superiority of NVP on the popular UVG benchmark; compared with prior arts, NVP not only trains 2 times faster (less than 5 minutes) but also exceeds their encoding quality as 34.07$\rightarrow$34.57 (measured with the PSNR metric), even using $>$8 times fewer parameters. We also show intriguing properties of NVP, e.g., video inpainting, video frame interpolation, etc.

Multi-modal data abounds in biomedicine, such as radiology images and reports. Interpreting this data at scale is essential for improving clinical care and accelerating clinical research. Biomedical text with its complex semantics poses additional challenges in vision--language modelling compared to the general domain, and previous work has used insufficiently adapted models that lack domain-specific language understanding. In this paper, we show that principled textual semantic modelling can substantially improve contrastive learning in self-supervised vision--language processing. We release a language model that achieves state-of-the-art results in radiology natural language inference through its improved vocabulary and novel language pretraining objective leveraging semantics and discourse characteristics in radiology reports. Further, we propose a self-supervised joint vision--language approach with a focus on better text modelling. It establishes new state of the art results on a wide range of publicly available benchmarks, in part by leveraging our new domain-specific language model. We release a new dataset with locally-aligned phrase grounding annotations by radiologists to facilitate the study of complex semantic modelling in biomedical vision--language processing. A broad evaluation, including on this new dataset, shows that our contrastive learning approach, aided by textual-semantic modelling, outperforms prior methods in segmentation tasks, despite only using a global-alignment objective.

Montreal-based centre unites strengths of McGill University, ÉTS, Mila, CNRS, Université Paris-Saclay, and CentraleSupélec A consortium of research organizations has gathered together to form a new International Research Laboratory (IRL) focused on artificial intelligence (AI) in Montreal. The new centre gathers together McGill University, École de technologie supérieure (ÉTS), Mila – Quebec AI Institute, France’s Centre Nationale de la Recherche Scientifique (CNRS), Université Paris-Saclay, and the École CentraleSupélec. The move confirms Montreal’s status as a leader in AI. While great strides have been made in AI recently, there is still a pressing need for new theoretical knowledge to better understand not only the capacities of this new technology, but how it achieves its results. The ILLS will focus on five main themes of research: fundamental aspects of artificial intelligence, sequential (real-time) machine learning, robust autonomous systems, natural language and speech processing, and applications to computer vision, signals, and information processing. In addition, the new centre will emphasize interdisciplinary collaborations with an aim to develop new methodologies and integrate these techniques into learning systems. “This new laboratory confirms Montreal’s global leadership in AI,” said Benoit Boulet, Associate Vice-Principal, Research & Innovation at McGill University. “This is a major hub with a talent pool that continues to deepen, and McGill researchers and students are embedded at every level of this activity. This new initiative will offer opportunities for our researchers to make even more breakthrough discoveries.” “The expertise of ÉTS in AI includes several laboratories and research chairs in artificial intelligence. This collaboration between France and Quebec makes it possible to innovate and deepen research in AI, a cross-cutting discipline from which we can benefit in many fields, including health, the built environment, robotics, and the Internet of Things. It is therefore with pride that ÉTS welcomes the new ILLS centre within its establishment,” said Christian Casanova, Director of Research and Partnerships at ÉTS. “Through its tools of international cooperation, CNRS supports the most promising cutting-edge joint research projects. The new international research laboratory brings together a powerful network of researchers from France and Québec to advance the knowledge and applications of AI. For the CNRS, this new lab is also an opportunity to strengthen more broadly its ties with the whole Canadian AI community,” said Antoine Petit, Chairman and CEO of CNRS. “AI at Paris-Saclay involves nearly 1,000 researchers, teacher-researchers, engineers and technicians and around forty laboratories, grouped together within our DataIA Institute. We will make our contribution to the ILLS in the form of the mobility of researchers, including the reception of Canadian colleagues at Paris-Saclay, the reception of Masters trainees, thesis funding in particular/among others. The University of Paris-Saclay is honored and proud to be associated with this signing ceremony for the creation of the IRL ILLS and to ensure its joint supervision" added Michel Guidal, Deputy Vice-President Research Sciences and Engineering at Université Paris-Saclay. “The ILLS, resulting from an unprecedented and international union, offers a unique potential for progress in the field of AI. It is an honor for CentraleSupélec to participate with our prestigious partners in this laboratory. Backed by this research, our teaching will thus be at the forefront of the world in terms of AI,” added Romain Soubeyran, Director of CentraleSupélec. The ILLS will join a burgeoning artificial intelligence (AI) sector in Montreal, which has attracted other major investments from government and business for the past several years. As a result, the city is one of the world’s leading hubs in this domain, with an estimated 27,000 workers in AI-related technologies and over 14,000 post-secondary students enrolled in AI-related study programs. The ILLS is the latest such laboratory to be launched in Canada, specifically in Quebec. In 2014, the CNRS and the Fonds de recherche du Québec – Nature et technologie (FRQNT) signed a letter of intent to support and promote the tradition of scientific cooperation that exists between France and Quebec. This collaboration has resulted in two International Research Laboratories in Quebec, as well as other shared research activities across the province. The CNRS has also established three other IRLs in Canada in partnership with other institutions. Present at the signing ceremony were: Frédéric Sanchez (Consul General of France), Remi Quirion (Quebec’s Chief Scientist), Antoine Petit (CNRS), Suzanne Fortier (McGill University), Francois Gagnon (ETS), Michel Guidal (Université Paris-Saclay), Franck Richecoeur (École CentraleSupélec), and Laurence Beaulieu (Mila). About McGill University Founded in Montreal, Quebec, in 1821, McGill University is Canada’s top ranked medical doctoral university. McGill is consistently ranked as one of the top universities, both nationally and internationally. It is a world-renowned institution of higher learning with research activities spanning three campuses, 11 faculties, 13 professional schools, 300 programs of study and over 39,000 students, including more than 10,400 graduate students. McGill attracts students from over 150 countries around the world, its 12,000 international students making up 30% of the student body. Over half of McGill students claim a first language other than English, including approximately 20% of our students who say French is their mother tongue.

Quantifying uncertainty in predictions or, more generally, estimating the posterior conditional distribution, is a core challenge in machine learning and statistics. We introduce Convex Nonparanormal Regression (CNR), a conditional nonparanormal approach for coping with this task. CNR involves a convex optimization of a posterior defined via a rich dictionary of pre-defined non linear transformations on Gaussians. It can fit an arbitrary conditional distribution, including multimodal and non-symmetric posteriors. For the special but powerful case of a piecewise linear dictionary, we provide a closed form of the posterior mean which can be used for point-wise predictions. Finally, we demonstrate the advantages of CNR over classical competitors using synthetic and real world data.

International communities of invention and innovation, IFIP WG 9.7 International conference on the history of computing, HC 2016, Brooklyn, New York, May 25–29 2016, revised selected papers, Springer International Publishing AG Switzerland, Cham 2016, pages 181–185