Automatic music transcription is considered to be one of the hardest problems in music information retrieval, yet recent deep learning approaches have achieved substantial improvements on transcription performance. These approaches commonly employ supervised learning models that predict various time-frequency representations, by minimizing element-wise losses such as the cross entropy function. However, applying the loss in this manner assumes conditional independence of each label given the input, and thus cannot accurately express inter-label dependencies. To address this issue, we introduce an adversarial training scheme that operates directly on the time-frequency representations and makes the output distribution closer to the ground-truth. Through adversarial learning, we achieve a consistent improvement in both frame-level and note-level metrics over Onsets and Frames, a state-of-the-art music transcription model. Our results show that adversarial learning can significantly reduce the error rate while increasing the confidence of the model estimations. Our approach is generic and applicable to any transcription model based on multi-label predictions, which are very common in music signal analysis.

The need for labour intensive pixel-wise annotation is a major limitation of many fully supervised learning methods for image segmentation. In this paper, we propose a deep convolutional neural network for multi-class segmentation that circumvents this problem by being trainable on coarse data labels combined with only a very small number of images with pixel-wise annotations. We call this new labelling strategy 'lazy' labels. Image segmentation is then stratified into three connected tasks: rough detection of class instances, separation of wrongly connected objects without a clear boundary, and pixel-wise segmentation to find the accurate boundaries of each object. These problems are integrated into a multitask learning framework and the model is trained end-to-end in a semi-supervised fashion. The method is applied on a dataset of food microscopy images. We show that the model gives accurate segmentation results even if exact boundary labels are missing for a majority of the annotated data. This allows more flexibility and efficiency for training deep neural networks that are data hungry in a practical setting where manual annotation is expensive, by collecting more lazy (rough) annotations than precisely segmented images.

Polysomnography (PSG) is the gold standard for diagnosing sleep obstructive apnea (OSA). It allows monitoring of breathing events throughout the night. The detection of these events is usually done by trained sleep experts. However, this task is tedious, highly time-consuming and subject to important inter-scorer variability. In this study, we adapted our state-of-the-art deep learning method for sleep event detection, DOSED, to the detection of sleep breathing events in PSG for the diagnosis of OSA. We used a dataset of 52 PSG recordings with apnea-hypopnea event scoring from 5 trained sleep experts. We assessed the performance of the automatic approach and compared it to the inter-scorer performance for both the diagnosis of OSA severity and, at the microscale, for the detection of single breathing events. We observed that human sleep experts reached an average accuracy of 75\% while the automatic approach reached 81\% for sleep apnea severity diagnosis. The F1 score for individual event detection was 0.55 for experts and 0.57 for the automatic approach, on average. These results demonstrate that the automatic approach can perform at a sleep expert level for the diagnosis of OSA.

Neuroimaging experiments produce a large volume (gigabytes) of high-dimensional spatio-temporal data for a small number of sampled participants and stimuli. Analyses of this data commonly compute averages over all trials, ignoring variation within groups of participants and stimuli. To enable the analysis of fMRI data without this implicit assumption of uniformity, we propose Neural Topographic Factor Analysis (NTFA), a deep generative model that parameterizes factors as functions of embeddings for participants and stimuli. We evaluate NTFA on a synthetically generated dataset as well as on three datasets from fMRI experiments. Our results demonstrate that NTFA yields more accurate reconstructions than a state-of-the-art method with fewer parameters. Moreover, learned embeddings uncover latent categories of participants and stimuli, which suggests that NTFA takes a first step towards reasoning about individual variation in fMRI experiments.

We present Placeto, a reinforcement learning (RL) approach to efficiently find device placements for distributed neural network training. Unlike prior approaches that only find a device placement for a specific computation graph, Placeto can learn generalizable device placement policies that can be applied to any graph. We propose two key ideas in our approach: (1) we represent the policy as performing iterative placement improvements, rather than outputting a placement in one shot; (2) we use graph embeddings to capture relevant information about the structure of the computation graph, without relying on node labels for indexing. These ideas allow Placeto to train efficiently and generalize to unseen graphs. Our experiments show that Placeto requires up to 6.1x fewer training steps to find placements that are on par with or better than the best placements found by prior approaches. Moreover, Placeto is able to learn a generalizable placement policy for any given family of graphs, which can then be used without any retraining to predict optimized placements for unseen graphs from the same family. This eliminates the large overhead incurred by prior RL approaches whose lack of generalizability necessitates re-training from scratch every time a new graph is to be placed.

Representational similarity analysis (RSA) has been shown to be an effective framework to characterize brainactivity The scarcity of methods to characterize the representational profiles and deep neural network activations as dynamics creates a major barrier to answer interesting representational geometry by computing the pairwise questions such as: how are objects represented in the brain distances of the response patterns as a representational over the time course from early perception to categorical decision dissimilarity matrix (RDM). However, how to properly analyze making, does the object identification or visual categorization and visualize the representational geometry as dynamics follows a hierarchical classification paradigm; do different over the time course from stimulus onset to offset classes of objects merge and branch at different time is not well understood. In this work, we formulated points based on different tasks or recurrence paradigm; are the pipeline to understand representational dynamics these representational dynamics oscillatory or recurrent?

Global routing has been a historically challenging problem in electronic circuit design, where the challenge is to connect a large and arbitrary number of circuit components with wires without violating the design rules for the printed circuit boards or integrated circuits. Similar routing problems also exist in the design of complex hydraulic systems, pipe systems and logistic networks. Existing solutions typically consist of greedy algorithms and hard-coded heuristics. As such, existing approaches suffer from a lack of model flexibility and non-optimum solutions. As an alternative approach, this work presents a deep reinforcement learning method for solving the global routing problem in a simulated environment. At the heart of the proposed method is deep reinforcement learning that enables an agent to produce an optimal policy for routing based on the variety of problems it is presented with leveraging the conjoint optimization mechanism of deep reinforcement learning. Conjoint optimization mechanism is explained and demonstrated in details; the best network structure and the parameters of the learned model are explored. Based on the fine-tuned model, routing solutions and rewards are presented and analyzed. The results indicate that the approach can outperform the benchmark method of a sequential A* method, suggesting a promising potential for deep reinforcement learning for global routing and other routing or path planning problems in general. Another major contribution of this work is the development of a global routing problem sets generator with the ability to generate parameterized global routing problem sets with different size and constraints, enabling evaluation of different routing algorithms and the generation of training datasets for future data-driven routing approaches.

"Theorem proving is similar to the game of Go. So, we can probably improve our provers using deep learning, like DeepMind built the super-human computer Go program, AlphaGo." Such optimism has been observed among participants of AITP2017. But is theorem proving really similar to Go? In this paper, we first identify the similarities and differences between them and then propose a system in which various provers keep competing against each other and changing themselves until they prove conjectures provided by users.

We consider the task of classifying when an extremely reduced amount of labelled data is available. This problem is of a great interest, in several real-world problems, as obtaining large amounts of labelled data is expensive and time consuming. We present a novel semi-supervised framework for multi-class classification that is based on the normalised and non-smooth graph 1-Laplacian. Our transductive framework is framed under a novel functional with carefully selected class priors - that enforces a sufficiently smooth solution that strengthens the intrinsic relation between the labelled and unlabelled data. We demonstrate through extensive experimental results on large datasets CIFAR-10 and ChestX-ray14, that our method outperforms classic methods and readily competes with recent deep-learning approaches.

An active new area in medicine involves training deep-learning models to detect structural patterns in brain scans associated with neurological diseases and disorders, such as Alzheimer's disease and multiple sclerosis. But collecting the training data is laborious: All anatomical structures in each scan must be separately outlined or hand-labeled by neurological experts. And, in some cases, such as for rare brain conditions in children, only a few scans may be available in the first place. In a paper presented at the recent Conference on Computer Vision and Pattern Recognition, the MIT researchers describe a system that uses a single labeled scan, along with unlabeled scans, to automatically synthesize a massive dataset of distinct training examples. The dataset can be used to better train machine-learning models to find anatomical structures in new scans -- the more training data, the better those predictions.