In many imaging modalities, objects of interest can occur in a variety of locations and poses (i.e. are subject to translations and rotations in 2d or 3d), but the location and pose of an object does not change its semantics (i.e. the object's essence). That is, the specific location and rotation of an airplane in satellite imagery, or the 3d rotation of a chair in a natural image, or the rotation of a particle in a cryo-electron micrograph, do not change the intrinsic nature of those objects. Here, we consider the problem of learning semantic representations of objects that are invariant to pose and location in a fully unsupervised manner. We address shortcomings in previous approaches to this problem by introducing TARGET-VAE, a translation and rotation group-equivariant variational autoencoder framework. TARGET-VAE combines three core innovations: 1) a rotation and translation group-equivariant encoder architecture, 2) a structurally disentangled distribution over latent rotation, translation, and a rotation-translation-invariant semantic object representation, which are jointly inferred by the approximate inference network, and 3) a spatially equivariant generator network. In comprehensive experiments, we show that TARGET-VAE learns disentangled representations without supervision that significantly improve upon, and avoid the pathologies of, previous methods. When trained on images highly corrupted by rotation and translation, the semantic representations learned by TARGET-VAE are similar to those learned on consistently posed objects, dramatically improving clustering in the semantic latent space. Furthermore, TARGET-VAE is able to perform remarkably accurate unsupervised pose and location inference. We expect methods like TARGET-VAE will underpin future approaches for unsupervised object generation, pose prediction, and object detection. Our code is available at https://github.com/SMLC-NYSBC/TARGET-VAE.

Human leukocyte antigen (HLA) complex molecules play an essential role in immune interactions by presenting peptides on the cell surface to T cells. With significant deep learning progress, a series of neural network-based models have been proposed and demonstrated with their excellent performances for peptide-HLA class I binding prediction. However, there is still a lack of effective binding prediction models for HLA class II protein binding with peptides due to its inherent challenges. We present a novel sequence-based pan-specific neural network structure, DeepSeaPanII, for peptide-HLA class II binding prediction in this work. Our model is an end-to-end neural network model without the need for pre-or post-processing on input samples compared with existing pan-specific models. Besides state-of-the-art performance in binding affinity prediction, DeepSeqPanII can also extract biological insight on the binding mechanism over the peptide by its attention mechanism-based binding core prediction capability. The leave-one-allele-out cross-validation and benchmark evaluation results show that our proposed network model achieved state-of-the-art performance in HLA-II peptide binding. The source code and trained models are freely available at https://github.com/pcpLiu/DeepSeqPanII.

Deep learning methods have recently made significant contributions to sound event detection. These methods either use a block-level approach to distinguish parts of audio containing the event, or analyze the small frames of the audio separately. In this paper, we introduce a new method, AudioMask, for rare sound event detection by combining these two approaches. AudioMask first applies Mask R-CNN, a state-of-the-art algorithm for detecting objects in images, to the log mel-spectrogram of the audio files. Mask R-CNN detects audio segments that might contain the target event by generating bounding boxes around them in time-frequency domain. Then we use a frame-based audio event classifier trained independently from Mask R-CNN, to analyze each individual frame in the candidate segments proposed by Mask R-CNN. A post-processing step combines the outputs of the Mask R-CNN and the frame-level classifier to identify the true events. By evaluating AudioMask over the data sets from 2017 Detection and Classification of Acoustic Scenes and Events (DCASE) Challenge Task 2, We show that our algorithm performs better than the baseline models by 13.3% in the average F-score and achieves better results compared to the other non-ensemble methods in the challenge.

This paper proposes an H∞robust state-feedback controller design for uncertain Takagi-Sugeno fuzzy systems using a nonmonotonic Lyapunov function. In the nonmonotonic approach, the monotonicity requirement of the Lyapunov function is relaxed by allowing it to increase locally. Based on the nonmonotonic Lyapunov function approach, sufficient conditions for the existence of a robust state-feedback H∞controller that guarantees stability and a prescribed H∞performance are given in terms of linear matrix inequalities. The proposed design technique is shown to be less conservative than the existing k-samples variations of the Lyapunov function. The effectiveness of the proposed approach is further illustrated via numerical examples.

This paper investigates the problems of robust H∞control for a class of discrete-time nonhomogenous Markovian jump linear systems (NMJLSs) by a multistep Lyapunov function (LF) approach. The proposed multistep LF is allowed to increase during the period of several sampling time step ahead of the current time within the Jump mode. First, a less conservative stability criterion is derived based on this multistep LF approach. Second, an H∞performance is analyzed under the multistep case by properly dealing with the knowledge of future states and exogenous noises. These two results are then employed to facilitate a robust H∞control design for NMJLSs, which yields a better disturbance attenuation performance as compared with the existing results. Numerical examples are included to elucidate the effectiveness of the developed results.

This paper recasts the problem of online secondary path modeling in the form of a statistical inverse problem. A statistical and, in particular, a Bayesian approach towards secondary path modeling is developed and the computational issues that emerge from this approach are discussed. All signals and parameters are modeled as random variables and the degree of information concerning them is coded in their probability density functions. An abstract solution is formulated in the form of a probability density function for the secondary path model. For extracting point estimates, common statistical estimation methods are investigated. It is shown that maximum likelihood estimation is not stable; however, Bayesian method of maximum a posteriori gives a reliable solution. An adaptive algorithm is then developed to compute this solution in a computationally efficient manner. This algorithm has three advantages, compared to the traditional secondary path modeling algorithms. First, it does not cause any interference with the main active noise control algorithm. Second, it does not require any additive-noise to be injected into the secondary path. Third, it does not require any off-line initiation. The convergence of the proposed algorithm is analyzed theoretically. The validity of the theoretical results is investigated by using computer simulation. Finally successful integration of the proposed algorithm into a real-time ANC system is reported.