09,24,2021

 Acting upon Imagination: when to trust imagined trajectories in model based reinforcement learning2021-05-13   ${\displaystyle \cong }$ Model based reinforcement learning (MBRL) uses an imperfect model of the world to imagine trajectories of future states and plan the best actions to maximize a reward function. These trajectories are imperfect and MBRL attempts to overcome this by relying on model predictive control (MPC) to continuously re-imagine trajectories from scratch. Such re-generation of imagined trajectories carries the major computational cost and increasing complexity in tasks with longer receding horizon. This paper aims to investigate how far in the future the imagined trajectories can be relied upon while still maintaining acceptable reward. Firstly, an error analysis is presented for systematic skipping recalculations for varying number of consecutive steps.% in several challenging benchmark control tasks. Secondly, we propose two methods offering when to trust and act upon imagined trajectories, looking at recent errors with respect to expectations, or comparing the confidence in an action imagined against its execution. Thirdly, we evaluate the effects of acting upon imagination while training the model of the world. Results show that acting upon imagination can reduce calculations by at least 20% and up to 80%, depending on the environment, while retaining acceptable reward. Using Self-Supervised Co-Training to Improve Facial Representation2021-05-13   ${\displaystyle \cong }$ In this paper, at first, the impact of ImageNet pre-training on Facial Expression Recognition (FER) was tested under different augmentation levels. It could be seen from the results that training from scratch could reach better performance compared to ImageNet fine-tuning at stronger augmentation levels. After that, a framework was proposed for standard Supervised Learning (SL), called Hybrid Learning (HL) which used Self-Supervised co-training with SL in Multi-Task Learning (MTL) manner. Leveraging Self-Supervised Learning (SSL) could gain additional information from input data like spatial information from faces which helped the main SL task. It is been investigated how this method could be used for FER problems with self-supervised pre-tasks such as Jigsaw puzzling and in-painting. The supervised head (SH) was helped by these two methods to lower the error rate under different augmentations and low data regime in the same training settings. The state-of-the-art was reached on AffectNet via two completely different HL methods, without utilizing additional datasets. Moreover, HL's effect was shown on two different facial-related problem, head poses estimation and gender recognition, which concluded to reduce in error rate by up to 9% and 1% respectively. Also, we saw that the HL methods prevented the model from reaching overfitting. SyntheticFur dataset for neural rendering2021-05-13   ${\displaystyle \cong }$ We introduce a new dataset called SyntheticFur built specifically for machine learning training. The dataset consists of ray traced synthetic fur renders with corresponding rasterized input buffers and simulation data files. We procedurally generated approximately 140,000 images and 15 simulations with Houdini. The images consist of fur groomed with different skin primitives and move with various motions in a predefined set of lighting environments. We also demonstrated how the dataset could be used with neural rendering to significantly improve fur graphics using inexpensive input buffers by training a conditional generative adversarial network with perceptual loss. We hope the availability of such high fidelity fur renders will encourage new advances with neural rendering for a variety of applications. Audio Captioning with Composition of Acoustic and Semantic Information2021-05-13   ${\displaystyle \cong }$ Generating audio captions is a new research area that combines audio and natural language processing to create meaningful textual descriptions for audio clips. To address this problem, previous studies mostly use the encoder-decoder based models without considering semantic information. To fill this gap, we present a novel encoder-decoder architecture using bi-directional Gated Recurrent Units (BiGRU) with audio and semantic embeddings. We extract semantic embedding by obtaining subjects and verbs from the audio clip captions and combine these embedding with audio embedding to feed the BiGRU-based encoder-decoder model. To enable semantic embeddings for the test audios, we introduce a Multilayer Perceptron classifier to predict the semantic embeddings of those clips. We also present exhaustive experiments to show the efficiency of different features and datasets for our proposed model the audio captioning task. To extract audio features, we use the log Mel energy features, VGGish embeddings, and a pretrained audio neural network (PANN) embeddings. Extensive experiments on two audio captioning datasets Clotho and AudioCaps show that our proposed model outperforms state-of-the-art audio captioning models across different evaluation metrics and using the semantic information improves the captioning performance. Keywords: Audio captioning; PANNs; VGGish; GRU; BiGRU. MapGo: Model-Assisted Policy Optimization for Goal-Oriented Tasks2021-05-13   ${\displaystyle \cong }$ In Goal-oriented Reinforcement learning, relabeling the raw goals in past experience to provide agents with hindsight ability is a major solution to the reward sparsity problem. In this paper, to enhance the diversity of relabeled goals, we develop FGI (Foresight Goal Inference), a new relabeling strategy that relabels the goals by looking into the future with a learned dynamics model. Besides, to improve sample efficiency, we propose to use the dynamics model to generate simulated trajectories for policy training. By integrating these two improvements, we introduce the MapGo framework (Model-Assisted Policy Optimization for Goal-oriented tasks). In our experiments, we first show the effectiveness of the FGI strategy compared with the hindsight one, and then show that the MapGo framework achieves higher sample efficiency when compared to model-free baselines on a set of complicated tasks. Graph Learning based Recommender Systems: A Review2021-05-13   ${\displaystyle \cong }$ Recent years have witnessed the fast development of the emerging topic of Graph Learning based Recommender Systems (GLRS). GLRS employ advanced graph learning approaches to model users' preferences and intentions as well as items' characteristics for recommendations. Differently from other RS approaches, including content-based filtering and collaborative filtering, GLRS are built on graphs where the important objects, e.g., users, items, and attributes, are either explicitly or implicitly connected. With the rapid development of graph learning techniques, exploring and exploiting homogeneous or heterogeneous relations in graphs are a promising direction for building more effective RS. In this paper, we provide a systematic review of GLRS, by discussing how they extract important knowledge from graph-based representations to improve the accuracy, reliability and explainability of the recommendations. First, we characterize and formalize GLRS, and then summarize and categorize the key challenges and main progress in this novel research area. Finally, we share some new research directions in this vibrant area. Explainable Machine Learning for Fraud Detection2021-05-13   ${\displaystyle \cong }$ The application of machine learning to support the processing of large datasets holds promise in many industries, including financial services. However, practical issues for the full adoption of machine learning remain with the focus being on understanding and being able to explain the decisions and predictions made by complex models. In this paper, we explore explainability methods in the domain of real-time fraud detection by investigating the selection of appropriate background datasets and runtime trade-offs on both supervised and unsupervised models. HeunNet: Extending ResNet using Heun's Methods2021-05-13   ${\displaystyle \cong }$ There is an analogy between the ResNet (Residual Network) architecture for deep neural networks and an Euler solver for an ODE. The transformation performed by each layer resembles an Euler step in solving an ODE. We consider the Heun Method, which involves a single predictor-corrector cycle, and complete the analogy, building a predictor-corrector variant of ResNet, which we call a HeunNet. Just as Heun's method is more accurate than Euler's, experiments show that HeunNet achieves high accuracy with low computational (both training and test) time compared to both vanilla recurrent neural networks and other ResNet variants. SAFIN: Arbitrary Style Transfer With Self-Attentive Factorized Instance Normalization2021-05-13   ${\displaystyle \cong }$ Artistic style transfer aims to transfer the style characteristics of one image onto another image while retaining its content. Existing approaches commonly leverage various normalization techniques, although these face limitations in adequately transferring diverse textures to different spatial locations. Self-Attention-based approaches have tackled this issue with partial success but suffer from unwanted artifacts. Motivated by these observations, this paper aims to combine the best of both worlds: self-attention and normalization. That yields a new plug-and-play module that we nameSelf-Attentive Fac-torized Instance Normalization(SAFIN). SAFIN is essentially a spatially adaptive normalization module whose parameters are inferred through attention on the content and style image. We demonstrate that plugging SAFIN into the base network of another state-of-the-art method results in enhanced stylization. We also develop a novel base network composed of Wavelet Transform for multi-scale style transfer, which when combined with SAFIN, produces visually appealing results with lesser unwanted textures. Good and Bad Optimization Models: Insights from Rockafellians2021-05-13   ${\displaystyle \cong }$ A basic requirement for a mathematical model is often that its solution (output) shouldn't change much if the model's parameters (input) are perturbed. This is important because the exact values of parameters may not be known and one would like to avoid being mislead by an output obtained using incorrect values. Thus, it's rarely enough to address an application by formulating a model, solving the resulting optimization problem and presenting the solution as the answer. One would need to confirm that the model is suitable, i.e., "good," and this can, at least in part, be achieved by considering a family of optimization problems constructed by perturbing parameters of concern. The resulting sensitivity analysis uncovers troubling situations with unstable solutions, which we referred to as "bad" models, and indicates better model formulations. Embedding an actual problem of interest within a family of problems is also a primary path to optimality conditions as well as computationally attractive, alternative problems, which under ideal circumstances, and when properly tuned, may even furnish the minimum value of the actual problem. The tuning of these alternative problems turns out to be intimately tied to finding multipliers in optimality conditions and thus emerges as a main component of several optimization algorithms. In fact, the tuning amounts to solving certain dual optimization problems. In this tutorial, we'll discuss the opportunities and insights afforded by this broad perspective. Informed Equation Learning2021-05-13   ${\displaystyle \cong }$ Distilling data into compact and interpretable analytic equations is one of the goals of science. Instead, contemporary supervised machine learning methods mostly produce unstructured and dense maps from input to output. Particularly in deep learning, this property is owed to the generic nature of simple standard link functions. To learn equations rather than maps, standard non-linearities can be replaced with structured building blocks of atomic functions. However, without strong priors on sparsity and structure, representational complexity and numerical conditioning limit this direct approach. To scale to realistic settings in science and engineering, we propose an informed equation learning system. It provides a way to incorporate expert knowledge about what are permitted or prohibited equation components, as well as a domain-dependent structured sparsity prior. Our system then utilizes a robust method to learn equations with atomic functions exhibiting singularities, as e.g. logarithm and division. We demonstrate several artificial and real-world experiments from the engineering domain, in which our system learns interpretable models of high predictive power. An Empirical Comparison of Bias Reduction Methods on Real-World Problems in High-Stakes Policy Settings2021-05-13   ${\displaystyle \cong }$ Applications of machine learning (ML) to high-stakes policy settings -- such as education, criminal justice, healthcare, and social service delivery -- have grown rapidly in recent years, sparking important conversations about how to ensure fair outcomes from these systems. The machine learning research community has responded to this challenge with a wide array of proposed fairness-enhancing strategies for ML models, but despite the large number of methods that have been developed, little empirical work exists evaluating these methods in real-world settings. Here, we seek to fill this research gap by investigating the performance of several methods that operate at different points in the ML pipeline across four real-world public policy and social good problems. Across these problems, we find a wide degree of variability and inconsistency in the ability of many of these methods to improve model fairness, but post-processing by choosing group-specific score thresholds consistently removes disparities, with important implications for both the ML research community and practitioners deploying machine learning to inform consequential policy decisions. Interval Deep Learning for Uncertainty Quantification in Safety Applications2021-05-13   ${\displaystyle \cong }$ Deep neural networks (DNNs) are becoming more prevalent in important safety-critical applications, where reliability in the prediction is paramount. Despite their exceptional prediction capabilities, current DNNs do not have an implicit mechanism to quantify and propagate significant input data uncertainty -- which is common in safety-critical applications. In many cases, this uncertainty is epistemic and can arise from multiple sources, such as lack of knowledge about the data generating process, imprecision, ignorance, and poor understanding of physics phenomena. Recent approaches have focused on quantifying parameter uncertainty, but approaches to end-to-end training of DNNs with epistemic input data uncertainty are more limited and largely problem-specific. In this work, we present a DNN optimized with gradient-based methods capable to quantify input and parameter uncertainty by means of interval analysis, which we call Deep Interval Neural Network (DINN). We perform experiments on an air pollution dataset with sensor uncertainty and show that the DINN can produce accurate bounded estimates from uncertain input data. BWCP: Probabilistic Learning-to-Prune Channels for ConvNets via Batch Whitening2021-05-13   ${\displaystyle \cong }$ This work presents a probabilistic channel pruning method to accelerate Convolutional Neural Networks (CNNs). Previous pruning methods often zero out unimportant channels in training in a deterministic manner, which reduces CNN's learning capacity and results in suboptimal performance. To address this problem, we develop a probability-based pruning algorithm, called batch whitening channel pruning (BWCP), which can stochastically discard unimportant channels by modeling the probability of a channel being activated. BWCP has several merits. (1) It simultaneously trains and prunes CNNs from scratch in a probabilistic way, exploring larger network space than deterministic methods. (2) BWCP is empowered by the proposed batch whitening tool, which is able to empirically and theoretically increase the activation probability of useful channels while keeping unimportant channels unchanged without adding any extra parameters and computational cost in inference. (3) Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet with various network architectures show that BWCP outperforms its counterparts by achieving better accuracy given limited computational budgets. For example, ResNet50 pruned by BWCP has only 0.70\% Top-1 accuracy drop on ImageNet, while reducing 43.1\% FLOPs of the plain ResNet50. Provably Convergent Algorithms for Solving Inverse Problems Using Generative Models2021-05-13   ${\displaystyle \cong }$ The traditional approach of hand-crafting priors (such as sparsity) for solving inverse problems is slowly being replaced by the use of richer learned priors (such as those modeled by deep generative networks). In this work, we study the algorithmic aspects of such a learning-based approach from a theoretical perspective. For certain generative network architectures, we establish a simple non-convex algorithmic approach that (a) theoretically enjoys linear convergence guarantees for certain linear and nonlinear inverse problems, and (b) empirically improves upon conventional techniques such as back-propagation. We support our claims with the experimental results for solving various inverse problems. We also propose an extension of our approach that can handle model mismatch (i.e., situations where the generative network prior is not exactly applicable). Together, our contributions serve as building blocks towards a principled use of generative models in inverse problems with more complete algorithmic understanding. Bootstrapping User and Item Representations for One-Class Collaborative Filtering2021-05-13   ${\displaystyle \cong }$ The goal of one-class collaborative filtering (OCCF) is to identify the user-item pairs that are positively-related but have not been interacted yet, where only a small portion of positive user-item interactions (e.g., users' implicit feedback) are observed. For discriminative modeling between positive and negative interactions, most previous work relied on negative sampling to some extent, which refers to considering unobserved user-item pairs as negative, as actual negative ones are unknown. However, the negative sampling scheme has critical limitations because it may choose "positive but unobserved" pairs as negative. This paper proposes a novel OCCF framework, named as BUIR, which does not require negative sampling. To make the representations of positively-related users and items similar to each other while avoiding a collapsed solution, BUIR adopts two distinct encoder networks that learn from each other; the first encoder is trained to predict the output of the second encoder as its target, while the second encoder provides the consistent targets by slowly approximating the first encoder. In addition, BUIR effectively alleviates the data sparsity issue of OCCF, by applying stochastic data augmentation to encoder inputs. Based on the neighborhood information of users and items, BUIR randomly generates the augmented views of each positive interaction each time it encodes, then further trains the model by this self-supervision. Our extensive experiments demonstrate that BUIR consistently and significantly outperforms all baseline methods by a large margin especially for much sparse datasets in which any assumptions about negative interactions are less valid. Paying Attention to Astronomical Transients: Photometric Classification with the Time-Series Transformer2021-05-13   ${\displaystyle \cong }$ Future surveys such as the Legacy Survey of Space and Time (LSST) of the Vera C. Rubin Observatory will observe an order of magnitude more astrophysical transient events than any previous survey before. With this deluge of photometric data, it will be impossible for all such events to be classified by humans alone. Recent efforts have sought to leverage machine learning methods to tackle the challenge of astronomical transient classification, with ever improving success. Transformers are a recently developed deep learning architecture, first proposed for natural language processing, that have shown a great deal of recent success. In this work we develop a new transformer architecture, which uses multi-head self attention at its core, for general multi-variate time-series data. Furthermore, the proposed time-series transformer architecture supports the inclusion of an arbitrary number of additional features, while also offering interpretability. We apply the time-series transformer to the task of photometric classification, minimising the reliance of expert domain knowledge for feature selection, while achieving results comparable to state-of-the-art photometric classification methods. We achieve a weighted logarithmic-loss of 0.507 on imbalanced data in a representative setting using data from the Photometric LSST Astronomical Time-Series Classification Challenge (PLAsTiCC). Moreover, we achieve a micro-averaged receiver operating characteristic area under curve of 0.98 and micro-averaged precision-recall area under curve of 0.87. DiscoBox: Weakly Supervised Instance Segmentation and Semantic Correspondence from Box Supervision2021-05-13   ${\displaystyle \cong }$ We introduce DiscoBox, a novel framework that jointly learns instance segmentation and semantic correspondence using bounding box supervision. Specifically, we propose a self-ensembling framework where instance segmentation and semantic correspondence are jointly guided by a structured teacher in addition to the bounding box supervision. The teacher is a structured energy model incorporating a pairwise potential and a cross-image potential to model the pairwise pixel relationships both within and across the boxes. Minimizing the teacher energy simultaneously yields refined object masks and dense correspondences between intra-class objects, which are taken as pseudo-labels to supervise the task network and provide positive/negative correspondence pairs for dense constrastive learning. We show a symbiotic relationship where the two tasks mutually benefit from each other. Our best model achieves 37.9% AP on COCO instance segmentation, surpassing prior weakly supervised methods and is competitive to supervised methods. We also obtain state of the art weakly supervised results on PASCAL VOC12 and PF-PASCAL with real-time inference. Feature Interactions on Steroids: On the Composition of ML Models2021-05-13   ${\displaystyle \cong }$ The lack of specifications is a key difference between traditional software engineering and machine learning. We discuss how it drastically impacts how we think about divide-and-conquer approaches to system design, and how it impacts reuse, testing and debugging activities. Traditionally, specifications provide a cornerstone for compositional reasoning and for the divide-and-conquer strategy of how we build large and complex systems from components, but those are hard to come by for machine-learned components. While the lack of specification seems like a fundamental new problem at first sight, in fact software engineers routinely deal with iffy specifications in practice: we face weak specifications, wrong specifications, and unanticipated interactions among components and their specifications. Machine learning may push us further, but the problems are not fundamentally new. Rethinking machine-learning model composition from the perspective of the feature interaction problem, we may even teach us a thing or two on how to move forward, including the importance of integration testing, of requirements engineering, and of design. OpenFL: An open-source framework for Federated Learning2021-05-13   ${\displaystyle \cong }$ Federated learning (FL) is a computational paradigm that enables organizations to collaborate on machine learning (ML) projects without sharing sensitive data, such as, patient records, financial data, or classified secrets. Open Federated Learning (OpenFL https://github.com/intel/openfl) is an open-source framework for training ML algorithms using the data-private collaborative learning paradigm of FL. OpenFL works with training pipelines built with both TensorFlow and PyTorch, and can be easily extended to other ML and deep learning frameworks. Here, we summarize the motivation and development characteristics of OpenFL, with the intention of facilitating its application to existing ML model training in a production environment. Finally, we describe the first use of the OpenFL framework to train consensus ML models in a consortium of international healthcare organizations, as well as how it facilitates the first computational competition on FL.