Nonlinear independent component analysis (ICA) is a general framework for unsupervised representation learning, and aimed at recovering the latent variables in data. Recent practical methods perform nonlinear ICA by solving a series of classification problems based on logistic regression. However, it is well-known that logistic regression is vulnerable to outliers, and thus the performance can be strongly weakened by outliers. In this paper, we first theoretically analyze nonlinear ICA models in the presence of outliers. Our analysis implies that estimation in nonlinear ICA can be seriously hampered when outliers exist on the tails of the (noncontaminated) target density, which happens in a typical case of contamination by outliers. We develop two robust nonlinear ICA methods based on the {\gamma}-divergence, which is a robust alternative to the KL-divergence in logistic regression. The proposed methods are shown to have desired robustness properties in the context of nonlinear ICA. We also experimentally demonstrate that the proposed methods are very robust and outperform existing methods in the presence of outliers. Finally, the proposed method is applied to ICA-based causal discovery and shown to find a plausible causal relationship on fMRI data.

--In recent years, huge amounts of unstructured textual data on the Internet are a big difficulty for AI algorithms to provide the best recommendations for users and their search queries. Since the Internet became widespread, a lot of research has been done in the field of Natural Language Processing (NLP) and machine learning. Almost every solution transforms documents into V ector Space Models (VSM) in order to apply AI algorithms over them. One such approach is based on Case-Based Reasoning (CBR). Therefore, the most important part of those systems is to compute the similarity between numerical data points. In 2016, the new similarity TS-SS metric is proposed, which showed state-of-the-art results in the field of textual mining for unsupervised learning. However, no one before has investigated its performances for supervised learning (classification task). In this work, we devised a CBR system capable of finding the most similar documents for a given query aiming to investigate performances of the new state-of- the-art metric, TS-SS, in addition to the two other geometrical similarity measures -- Euclidean distance and Cosine similarity -- that showed the best predictive results over several benchmark corpora. The results show surprising inappropriateness of TS-SS measure for high dimensional features.

Abstract--Federated learning is a communication-efficient training process that alternates between local training at the edge devices and averaging the updated local model at the central server . Nevertheless, it is impractical to achieve a perfect acquisition of the local models in wireless communication d ue to noise, which also brings serious effects on federated learn ing. T o tackle this challenge, we propose a robust design for federa ted learning to alleviate the effects of noise in this paper . Con sidering noise in the two aforementioned steps, we first formulate the training problem as a parallel optimization for each node un der the expectation-based model and the worst-case model. Due t o the non-convexity of the problem, a regularization for the l oss function approximation method is proposed to make it tracta ble. Regarding the worst-case model, we develop a feasible train ing scheme which utilizes the sampling-based successive conve x approximation algorithm to tackle the unavailable maxima o r minima noise condition and the non-convex issue of the objec tive function. Furthermore, the convergence rates of both new de signs are analyzed from a theoretical point of view. Finally, the improvement of prediction accuracy and the reduction of los s function are demonstrated via simulations for the proposed designs. UTURE wireless computing applications demand higher bandwidth, lower latency and more reliable connections with numerous devices [1].

Deep learning, in general, and natural language processing methods, in particular, rely heavily on annotated samples to achieve good performance. However, manually annotating data is expensive and time consuming. Active Learning (AL) strategies reduce the need for huge volumes of labelled data by iteratively selecting a small number of examples for manual annotation based on their estimated utility in training the given model. In this paper, we argue that since AL strategies choose examples independently, they may potentially select similar examples, all of which do not aid in the learning process. We propose a method, referred to as Active$\mathbf{^2}$ Learning (A$\mathbf{^2}$L), that actively adapts to the sequence tagging model being trained, to further eliminate such redundant examples chosen by an AL strategy. We empirically demonstrate that A$\mathbf{^2}$L improves the performance of state-of-the-art AL strategies on different sequence tagging tasks. Furthermore, we show that A$\mathbf{^2}$L is widely applicable by using it in conjunction with different AL strategies and sequence tagging models. We demonstrate that the proposed A$\mathbf{^2}$L able to reach full data F-score with $\approx\mathbf{2-16 \%}$ less data compared to state-of-art AL strategies on different sequence tagging datasets.

We consider the problem of aggregating models learned from sequestered, possibly heterogeneous datasets. Exploiting tools from Bayesian nonparametrics, we develop a general meta-modeling framework that learns shared global latent structures by identifying correspondences among local model parameterizations. Our proposed framework is model-independent and is applicable to a wide range of model types. After verifying our approach on simulated data, we demonstrate its utility in aggregating Gaussian topic models, hierarchical Dirichlet process based hidden Markov models, and sparse Gaussian processes with applications spanning text summarization, motion capture analysis, and temperature forecasting.

We consider the problem of distributed feature quantization, where the goal is to enable a pretrained classifier at a central node to carry out its classification on features that are gathered from distributed nodes through communication constrained channels. We propose the design of distributed quantization schemes specifically tailored to the classification task: unlike quantization schemes that help the central node reconstruct the original signal as accurately as possible, our focus is not reconstruction accuracy, but instead correct classification. Our work does not make any apriori distributional assumptions on the data, but instead uses training data for the quantizer design. Our main contributions include: we prove NP-hardness of finding optimal quantizers in the general case; we design an optimal scheme for a special case; we propose quantization algorithms, that leverage discrete neural representations and training data, and can be designed in polynomial-time for any number of features, any number of classes, and arbitrary division of features across the distributed nodes. We find that tailoring the quantizers to the classification task can offer significant savings: as compared to alternatives, we can achieve more than a factor of two reduction in terms of the number of bits communicated, for the same classification accuracy.

This paper addresses the challenging task of video captioning which aims to generate descriptions for video data. Recently, the attention-based encoder-decoder structures have been widely used in video captioning. In existing literature, the attention weights are often built from the information of an individual modality, while, the association relationships between multiple modalities are neglected. Motivated by this observation, we propose a video captioning model with High-Order Cross-Modal Attention (HOCA) where the attention weights are calculated based on the high-order correlation tensor to capture the frame-level cross-modal interaction of different modalities sufficiently. Furthermore, we novelly introduce Low-Rank HOCA which adopts tensor decomposition to reduce the extremely large space requirement of HOCA, leading to a practical and efficient implementation in real-world applications. Experimental results on two benchmark datasets, MSVD and MSR-VTT, show that Low-rank HOCA establishes a new state-of-the-art.

By analyzing Bayesian inference of generative model for random networks with both relations (edges) and node features (discrete labels), we perform an asymptotically exact analysis of the semi-supervised classfication problems on graph-structured data using the cavity method of statistical physics. We unveil detectability phase transitions which put fundamental limit on ability of classfications for all possible algorithms. Our theory naturally converts to a message passing algorithm which works all the way down to the phase transition in the underlying generative model, and can be translated to a graph convolution neural network algorithm which greatly outperforms existing algorithms including popular graph neural networks in synthetic networks. When applied to real-world datasets, our algorithm achieves comparable performance with the state-of-the art algorithms. Our approach provides benchmark datasets with continuously tunable parameters and optimal results, which can be used to evaluate performance of exiting graph neural networks, and to find and understand their strengths and limitations. In particular, we observe that popular GCNs have sparsity issue and ovefitting issue on large synthetic benchmarks, we also show how to overcome the issues by combining strengths of our approach.

Solving goal-oriented tasks is an important but challenging problem in reinforcement learning (RL). For such tasks, the rewards are often sparse, making it difficult to learn a policy effectively. To tackle this difficulty, we propose a new approach called Policy Continuation with Hindsight Inverse Dynamics (PCHID). This approach learns from Hindsight Inverse Dynamics based on Hindsight Experience Replay, enabling the learning process in a self-imitated manner and thus can be trained with supervised learning. This work also extends it to multi-step settings with Policy Continuation. The proposed method is general, which can work in isolation or be combined with other on-policy and off-policy algorithms. On two multi-goal tasks GridWorld and FetchReach, PCHID significantly improves the sample efficiency as well as the final performance.

The usage of deep generative models for image compression has led to impressive performance gains over classical codecs while neural video compression is still in its infancy. Here, we propose an end-to-end, deep generative modeling approach to compress temporal sequences with a focus on video. Our approach builds upon variational autoencoder (VAE) models for sequential data and combines them with recent work on neural image compression. The approach jointly learns to transform the original sequence into a lower-dimensional representation as well as to discretize and entropy code this representation according to predictions of the sequential VAE. Rate-distortion evaluations on small videos from public data sets with varying complexity and diversity show that our model yields competitive results when trained on generic video content. Extreme compression performance is achieved when training the model on specialized content.