**TensorFlow Eager: A Multi-Stage, Python-Embedded DSL for Machine Learning**

TensorFlow Eager is a multi-stage, Python-embedded domain-specific language for hardware-accelerated machine learning, suitable for both interactive research and production. TensorFlow, which TensorFlow Eager extends, requires users to represent computations as dataflow graphs; this permits compiler optimizations and simplifies deployment but hinders rapid prototyping and run-time dynamism. TensorFlow Eager eliminates these usability costs without sacrificing the benefits furnished by graphs: It provides an imperative front-end to TensorFlow that executes operations immediately and a JIT tracer that translates Python functions composed of TensorFlow operations into executable dataflow graphs. TensorFlow Eager thus offers a multi-stage programming model that makes it easy to interpolate between imperative and staged execution in a single package.

**Thinging for Computational Thinking**

This paper examines conceptual models and their application to computational thinking. Computational thinking is a fundamental skill for everybody, not just for computer scientists. It has been promoted as skills that are as fundamental for all as numeracy and literacy. According to authorities in the field, the best way to characterize computational thinking is the way in which computer scientists think and the manner in which they reason how computer scientists think for the rest of us. Core concepts in computational thinking include such notions as algorithmic thinking, abstraction, decomposition, and generalization. This raises several issues and challenges that still need to be addressed, including the fundamental characteristics of computational thinking and its relationship with modeling patterns (e.g., object-oriented) that lead to programming/coding. Thinking pattern refers to recurring templates used by designers in thinking. In this paper, we propose a representation of thinking activity by adopting a thinking pattern called thinging that utilizes a diagrammatic technique called thinging machine (TM). We claim that thinging is a valuable process as a fundamental skill for everybody in computational thinking. The viability of such a proclamation is illustrated through examples and a case study.

**Externalities in Knowledge Production: Evidence from a Randomized Field Experiment**

Are there positive or negative externalities in knowledge production? Do current contributions to knowledge production increase or decrease the future growth of knowledge? We use a randomized field experiment, which added relevant content to some pages in Wikipedia while leaving similar pages unchanged. We find that the addition of content has a negligible impact on the subsequent long-run growth of content. Our results have implications for information seeding and incentivizing contributions, implying that additional content does not generate sizable externalities by inspiring nor discouraging future contributions.

**Comparison of plotting system outputs in beginner analysts**

The R programming language is built on an ecosystem of packages, some that allow analysts to accomplish the same tasks. For example, there are at least two clear workflows for creating data visualizations in R: using the base graphics package (referred to as ‘base R’) and the ggplot2 add-on package based on the grammar of graphics. Here we perform an empirical study of the quality of scientific graphics produced by beginning R users. In our experiment, learners taking a data science course on the Coursera platform were randomized to complete identical plotting exercises in either the base R or the ggplot2 system. Learners were then asked to evaluate their peers in terms of visual characteristics key to scientific cognition. We observed that graphics created with the two systems rated similarly on many characteristics. However, ggplot2 graphics were generally judged to be more visually pleasing and, in the case of faceted scientific plots, easier to understand. Our results suggest that while both graphic systems are useful in the hands of beginning users, ggplot2’s natural faceting system may be easier to use by beginning users for displaying more complex relationships.

**Making the Dynamic Time Warping Distance Warping-Invariant**

The literature postulates that the dynamic time warping (dtw) distance can cope with temporal variations but stores and processes time series in a form as if the dtw-distance cannot cope with such variations. To address this inconsistency, we first show that the dtw-distance is not warping-invariant-despite its name and contrary to its characterization in some publications. The lack of warping-invariance contributes to the inconsistency mentioned above and to a strange behavior. To eliminate these peculiarities, we convert the dtw-distance to a warping-invariant semi-metric, called time-warp-invariant (twi) distance. Empirical results suggest that the error rates of the twi and dtw nearest-neighbor classifier are practically equivalent in a Bayesian sense. However, the twi-distance requires less storage and computation time than the dtw-distance for a broad range of problems. These results challenge the current practice of applying the dtw-distance in nearest-neighbor classification and suggest the proposed twi-distance as a more efficient and consistent option.

**The Fuzzy ROC**

The fuzzy ROC extends Receiver Operating Curve (ROC) visualization to the situation where some data points, falling in an indeterminacy region, are not classified. It addresses two challenges: definition of sensitivity and specificity bounds under indeterminacy; and visual summarization of the large number of possibilities arising from different choices of indeterminacy zones.

**Voyageur: An Experiential Travel Search Engine**

We describe Voyageur, which is an application of experiential search to the domain of travel. Unlike traditional search engines for online services, experiential search focuses on the experiential aspects of the service under consideration. In particular, Voyageur needs to handle queries for subjective aspects of the service (e.g., quiet hotel, friendly staff) and combine these with objective attributes, such as price and location. Voyageur also highlights interesting facts and tips about the services the user is considering to provide them with further insights into their choices.

**Modeling Social Group Communication with Multi-Agent Imitation Learning**

In crowded social scenarios with a myriad of external stimuli, human brains exhibit a natural ability to filter out irrelevant information and narrowly focus their attention. In the midst of multiple groups of people, humans use such sensory gating to effectively further their own group’s interactional goals. In this work, we consider the design of a policy network to model multi-group multi-person communication. Our policy takes as input the state of the world such as an agent’s gaze direction, body pose of other agents or history of past actions, and outputs an optimal action such as speaking, listening or responding (communication modes). Inspired by humans’ natural neurobiological filtering process, a central component of our policy network design is an information gating function, termed the Kinesic-Proxemic-Message Gate (KPM-Gate), that models the ability of an agent to selectively gather information from specific neighboring agents. The degree of influence of a neighbor is based on dynamic non-verbal cues such as body motion, head pose (kinesics) and interpersonal space (proxemics). We further show that the KPM-Gate can be used to discover social groups using its natural interpretation as a social attention mechanism. We pose the communication policy learning problem as a multi-agent imitation learning problem. We learn a single policy shared by all agents under the assumption of a decentralized Markov decision process. We term our policy network as the Multi-Agent Group Discovery and Communication Mode Network (MAGDAM network), as it learns social group structure in addition to the dynamics of group communication. Our experimental validation on both synthetic and real world data shows that our model is able to both discover social group structure and learn an accurate multi-agent communication policy.

**Model Primitive Hierarchical Lifelong Reinforcement Learning**

Learning interpretable and transferable subpolicies and performing task decomposition from a single, complex task is difficult. Some traditional hierarchical reinforcement learning techniques enforce this decomposition in a top-down manner, while meta-learning techniques require a task distribution at hand to learn such decompositions. This paper presents a framework for using diverse suboptimal world models to decompose complex task solutions into simpler modular subpolicies. This framework performs automatic decomposition of a single source task in a bottom up manner, concurrently learning the required modular subpolicies as well as a controller to coordinate them. We perform a series of experiments on high dimensional continuous action control tasks to demonstrate the effectiveness of this approach at both complex single task learning and lifelong learning. Finally, we perform ablation studies to understand the importance and robustness of different elements in the framework and limitations to this approach.

**Learning Dynamics Model in Reinforcement Learning by Incorporating the Long Term Future**

In model-based reinforcement learning, the agent interleaves between model learning and planning. These two components are inextricably intertwined. If the model is not able to provide sensible long-term prediction, the executed planner would exploit model flaws, which can yield catastrophic failures. This paper focuses on building a model that reasons about the long-term future and demonstrates how to use this for efficient planning and exploration. To this end, we build a latent-variable autoregressive model by leveraging recent ideas in variational inference. We argue that forcing latent variables to carry future information through an auxiliary task substantially improves long-term predictions. Moreover, by planning in the latent space, the planner’s solution is ensured to be within regions where the model is valid. An exploration strategy can be devised by searching for unlikely trajectories under the model. Our method achieves higher reward faster compared to baselines on a variety of tasks and environments in both the imitation learning and model-based reinforcement learning settings.

**A Nonparametric Dynamic Causal Model for Macroeconometrics**

This paper uses potential outcome time series to provide a nonparametric framework for quantifying dynamic causal effects in macroeconometrics. This provides sufficient conditions for the nonparametric identification of dynamic causal effects as well as clarify the causal content of several common assumptions and methods in macroeconomics. Our key identifying assumption is shown to be non-anticipating treatments which enables nonparametric inference on dynamic causal effects. Next, we provide a formal definition of a `shock’ and this leads to a shocked potential outcome time series. This is a nonparametric statement of the Frisch-Slutzky paradigm. The common additional assumptions that the causal effects are additive and that the treatments are shocks place substantial restrictions on the underlying dynamic causal estimands. We use this structure to causally interpret several common estimation strategies. We provide sufficient conditions under which local projections is causally interpretable and show that the standard assumptions for local projections with an instrument are not sufficient to identify dynamic causal effects. We finally show that the structural vector moving average form is causally equivalent to a restricted potential outcome time series under the usual invertibility assumption.

**Change Detection with the Kernel Cumulative Sum Algorithm**

Online change detection involves monitoring a stream of data for changes in the statistical properties of incoming observations. A good change detector will detect any changes shortly after they occur, while raising few false alarms. Although there are algorithms with confirmed optimality properties for this task, they rely on the exact specifications of the probability distributions involved and this limits their practicality. In this work we describe a kernel-based variant of the Cumulative Sum (CUSUM) change detection algorithm that can detect changes under less restrictive assumptions. Instead of using the likelihood ratio, which is a parametric quantity, the Kernel CUSUM (KCUSUM) algorithm relies on a statistic based on the Maximum Mean Discrepancy (MMD) non-parametric testing framework. The KCUSUM algorithm is applicable in settings where there is a large amount of background data available and it is desirable to detect a change away from this background setting. Exploiting the random-walk structure of the test statistic, we derive bounds on the performance of the algorithm, including the expected delay and the average time to false alarm.

**Causal Discovery and Hidden Driving Force Estimation from Nonstationary/Heterogeneous Data**

It is commonplace to encounter nonstationary or heterogeneous data. Such a distribution shift feature presents both challenges and opportunities for causal discovery, of which the underlying generating process changes over time or across domains. In this paper, we develop a principled framework for causal discovery from such data, called Constraint-based causal Discovery from NOnstationary/heterogeneous Data (CD-NOD), which addresses two important questions. First, we propose an enhanced constraint-based procedure to detect variables whose local mechanisms change and recover the skeleton of the causal structure over observed variables. Second, we present a way to determine causal orientations by making use of independent changes in the data distribution implied by the underlying causal model, benefiting from information carried by changing distributions. After learning the causal structure, next, we investigate how to efficiently estimate the `driving force’ of the nonstationarity of a causal mechanism. That is, we aim to extract from data a low-dimensional and interpretable representation of changes. The proposed methods are totally nonparametric, with no restrictions on data distributions and causal mechanisms, and do not rely on window segmentation. Furthermore, we find that nonstationarity benefits causal structure identification with particular types of confounders. Finally, we show the tight connection between nonstationarity/heterogeneity and soft intervention in causal discovery. Experimental results on various synthetic and real-world data sets (task-fMRI and stock data) are presented to demonstrate the efficacy of the proposed methods.

**Less is More: Semi-Supervised Causal Inference for Detecting Pathogenic Users in Social Media**

Recent years have witnessed a surge of manipulation of public opinion and political events by malicious social media actors. These users are referred to as ‘Pathogenic Social Media (PSM)’ accounts. PSMs are key users in spreading misinformation in social media to viral proportions. These accounts can be either controlled by real users or automated bots. Identification of PSMs is thus of utmost importance for social media authorities. The burden usually falls to automatic approaches that can identify these accounts and protect social media reputation. However, lack of sufficient labeled examples for devising and training sophisticated approaches to combat these accounts is still one of the foremost challenges facing social media firms. In contrast, unlabeled data is abundant and cheap to obtain thanks to massive user-generated data. In this paper, we propose a semi-supervised causal inference PSM detection framework, SemiPsm, to compensate for the lack of labeled data. In particular, the proposed method leverages unlabeled data in the form of manifold regularization and only relies on cascade information. This is in contrast to the existing approaches that use exhaustive feature engineering (e.g., profile information, network structure, etc.). Evidence from empirical experiments on a real-world ISIS-related dataset from Twitter suggests promising results of utilizing unlabeled instances for detecting PSMs.

**Measuring and Controlling Bias for Some Bayesian Inferences and the Relation to Frequentist Criteria**

A common concern with Bayesian methodology in scientific contexts is that inferences can be heavily influenced by subjective biases. As presented here, there are two types of bias for some quantity of interest: bias against and bias in favor. Based upon the principle of evidence, it is shown how to measure and control these biases for both hypothesis assessment and estimation problems. Optimality results are established for the principle of evidence as the basis of the approach to these problems. A close relationship is established between measuring bias in Bayesian inferences and frequentist properties that hold for any proper prior. This leads to a possible resolution to an apparent conflict between these approaches to statistical reasoning. Frequentism is seen as establishing a figure of merit for a statistical study, while Bayesianism plays the key role in determining inferences based upon statistical evidence.

**BOINC: A Platform for Volunteer Computing**

‘Volunteer computing’ is the use of consumer digital devices for high-throughput scientific computing. It can provide large computing capacity at low cost, but presents challenges due to device heterogeneity, unreliability, and churn. BOINC, a widely-used open-source middleware system for volunteer computing, addresses these challenges. We describe its features, architecture, and implementation.

**Tutorial: Deriving The Efficient Influence Curve for Large Models**

This paper aims to provide a tutorial for upper level undergraduate and graduate students in statistics and biostatistics on deriving influence functions for non-parametric and semi-parametric models. The author will build on previously known efficiency theory and provide a useful identity and formulaic technique only relying on the basics of integration which, are self-contained in this tutorial and can be used in most any setting one might encounter in practice. The paper provides many examples of such derivations for well-known influence functions as well as for new parameters of interest. The influence function remains a central object for constructing efficient estimators for large models, such as the one-step estimator and the targeted maximum likelihood estimator. We will not touch upon these estimators at all but readers familiar with these estimators might find this tutorial of particular use.

**Analysing Motifs in Multilayer Networks**

Network motifs can capture basic interaction patterns and inform the functional properties of networks. However, real-world complex systems often have multiple types of relationships, which cannot be represented by a monolayer network. The multilayer nature of complex systems demands research on extending the notion of motifs to multilayer networks, thereby exploring the interaction patterns with a higher resolution. In this paper, we propose a formal definition of multilayer motifs, and analyse the occurrence of three-node multilayer motifs in a set of real-world multilayer networks. We find that multilayer motifs in social networks are more homogeneous across layers, indicating that different types of social relationships are reinforcing each other, while those in the transportation network are more complementary across layers. We find that biological networks are often associated with heterogeneous functions. This research sheds light on how multilayer network framework enables the capture of the hidden multi-aspect relationships among the nodes.

**Probabilistic Modeling for Novelty Detection with Applications to Fraud Identification**

Novelty detection is the unsupervised problem of identifying anomalies in test data which significantly differ from the training set. Novelty detection is one of the classic challenges in Machine Learning and a core component of several research areas such as fraud detection, intrusion detection, medical diagnosis, data cleaning, and fault prevention. While numerous algorithms were designed to address this problem, most methods are only suitable to model continuous numerical data. Tackling datasets composed of mixed-type features, such as numerical and categorical data, or temporal datasets describing discrete event sequences is a challenging task. In addition to the supported data types, the key criteria for efficient novelty detection methods are the ability to accurately dissociate novelties from nominal samples, the interpretability, the scalability and the robustness to anomalies located in the training data. In this thesis, we investigate novel ways to tackle these issues. In particular, we propose (i) an experimental comparison of novelty detection methods for mixed-type data (ii) an experimental comparison of novelty detection methods for sequence data, (iii) a probabilistic nonparametric novelty detection method for mixed-type data based on Dirichlet process mixtures and exponential-family distributions and (iv) an autoencoder-based novelty detection model with encoder/decoder modelled as deep Gaussian processes.

**DeepStego: Protecting Intellectual Property of Deep Neural Networks by Steganography**

Deep Neural Networks (DNNs) has shown great success in various challenging tasks. Training these networks is computationally expensive and requires vast amounts of training data. Therefore, it is necessary to design a technology to protect the intellectual property (IP) of the model and externally verify the ownership of the model in a black-box way. Previous studies either fail to meet the black-box requirement or have not dealt with several forms of security and legal problems. In this paper, we firstly propose a novel steganographic scheme for watermarking Deep Neural Networks in the process of training. This scheme is the first feasible scheme to protect DNNs which perfectly solves the problems of safety and legality. We demonstrate experimentally that such a watermark has no obvious influence on the main task of model design and can successfully verify the ownership of the model. Furthermore, we show a rather robustness by simulating our scheme in a real situation.

**A New Approach to Adaptive Data Analysis and Learning via Maximal Leakage**

There is an increasing concern that most current published research findings are false. The main cause seems to lie in the fundamental disconnection between theory and practice in data analysis. While the former typically relies on statistical independence, the latter is an inherently adaptive process: new hypotheses are formulated based on the outcomes of previous analyses. A recent line of work tries to mitigate these issues by enforcing constraints, such as differential privacy, that compose adaptively while degrading gracefully and thus provide statistical guarantees even in adaptive contexts. Our contribution consists in the introduction of a new approach, based on the concept of Maximal Leakage, an information-theoretic measure of leakage of information. The main result allows us to compare the probability of an event happening when adaptivity is considered with respect to the non-adaptive scenario. The bound we derive represents a generalization of the bounds used in non-adaptive scenarios (e.g., McDiarmid’s inequality for

-sensitive functions, false discovery error control via significance level, etc.), and allows us to replicate or even improve, in certain regimes, the results obtained using Max-Information or Differential Privacy. In contrast with the line of work started by Dwork et al., our results do not rely on Differential Privacy but are, in principle, applicable to every algorithm that has a bounded leakage, including the differentially private algorithms and the ones with a short description length.

**Multiple-Kernel Dictionary Learning for Reconstruction and Clustering of Unseen Multivariate Time-series**

There exist many approaches for description and recognition of unseen classes in datasets. Nevertheless, it becomes a challenging problem when we deal with multivariate time-series (MTS) (e.g., motion data), where we cannot apply the vectorial algorithms directly to the inputs. In this work, we propose a novel multiple-kernel dictionary learning (MKD) which learns semantic attributes based on specific combinations of MTS dimensions in the feature space. Hence, MKD can fully/partially reconstructs the unseen classes based on the training data (seen classes). Furthermore, we obtain sparse encodings for unseen classes based on the learned MKD attributes, and upon which we propose a simple but effective incremental clustering algorithm to categorize the unseen MTS classes in an unsupervised way. According to the empirical evaluation of our MKD framework on real benchmarks, it provides an interpretable reconstruction of unseen MTS data as well as a high performance regarding their online clustering.

**Copying Machine Learning Classifiers**

We study model-agnostic copies of machine learning classifiers. We develop the theory behind the problem of copying, highlighting its differences with that of learning, and propose a framework to copy the functionality of any classifier using no prior knowledge of its parameters or training data distribution. We identify the different sources of loss and provide guidelines on how best to generate synthetic sets for the copying process. We further introduce a set of metrics to evaluate copies in practice. We validate our framework through extensive experiments using data from a series of well-known problems. We demonstrate the value of copies in use cases where desiderata such as interpretability, fairness or productivization constrains need to be addressed. Results show that copies can be exploited to enhance existing solutions and improve them adding new features and characteristics.

**Learning a smooth kernel regularizer for convolutional neural networks**

Modern deep neural networks require a tremendous amount of data to train, often needing hundreds or thousands of labeled examples to learn an effective representation. For these networks to work with less data, more structure must be built into their architectures or learned from previous experience. The learned weights of convolutional neural networks (CNNs) trained on large datasets for object recognition contain a substantial amount of structure. These representations have parallels to simple cells in the primary visual cortex, where receptive fields are smooth and contain many regularities. Incorporating smoothness constraints over the kernel weights of modern CNN architectures is a promising way to improve their sample complexity. We propose a smooth kernel regularizer that encourages spatial correlations in convolution kernel weights. The correlation parameters of this regularizer are learned from previous experience, yielding a method with a hierarchical Bayesian interpretation. We show that our correlated regularizer can help constrain models for visual recognition, improving over an L2 regularization baseline.

**Gated Graph Convolutional Recurrent Neural Networks**

Graph processes model a number of important problems such as identifying the epicenter of an earthquake or predicting weather. In this paper, we propose a Graph Convolutional Recurrent Neural Network (GCRNN) architecture specifically tailored to deal with these problems. GCRNNs use convolutional filter banks to keep the number of trainable parameters independent of the size of the graph and of the time sequences considered. We also put forward Gated GCRNNs, a time-gated variation of GCRNNs akin to LSTMs. When compared with GNNs and another graph recurrent architecture in experiments using both synthetic and real-word data, GCRNNs significantly improve performance while using considerably less parameters.

**BacSoft: A Tool to Archive Data on Bacteria**

Recently, DNA data storage systems have attracted many researchers worldwide. Motivated by the success stories of such systems, in this work we propose a software called BacSoft to clone the data in a bacterial plasmid by using the concept of genetic engineering. We consider the encoding schemes such that it satisfies constraints significant for bacterial data storage.

**Blockchain Meets Database: Design and Implementation of a Blockchain Relational Database**

In this paper, we design and implement the first-ever decentralized replicated relational database with blockchain properties that we term blockchain relational database. We highlight several similarities between features provided by blockchain platforms and a replicated relational database, although they are conceptually different, primarily in their trust model. Motivated by this, we leverage the rich features, decades of research and optimization, and available tooling in relational databases to build a blockchain relational database. We consider a permissioned blockchain model of known, but mutually distrustful organizations each operating their own database instance that are replicas of one another. The replicas execute transactions independently and engage in decentralized consensus to determine the commit order for transactions. We design two approaches, the first where the commit order for transactions is agreed upon prior to executing them, and the second where transactions are executed without prior knowledge of the commit order while the ordering happens in parallel. We leverage serializable snapshot isolation (SSI) to guarantee that the replicas across nodes remain consistent and respect the ordering determined by consensus, and devise a new variant of SSI based on block height for the latter approach. We implement our system on PostgreSQL and present detailed performance experiments analyzing both approaches.

**An approach to Decision Making based on Dynamic Argumentation Systems**

In this paper, we introduce a formalism for single-agent decision making that is based on Dynamic Argumentation Frameworks. The formalism can be used to justify a choice, which is based on the current situation the agent is involved. Taking advantage of the inference mechanism of the argumentation formalism, it is possible to consider preference relations and conflicts among the available alternatives for that reasoning. With this formalization, given a particular set of evidence, the justified conclusions supported by warranted arguments will be used by the agent’s decision rules to determine which alternatives will be selected. We also present an algorithm that implements a choice function based on our formalization. Finally, we complete our presentation by introducing formal results that relate the proposed framework with approaches of classical decision theory.

**O-GAN: Extremely Concise Approach for Auto-Encoding Generative Adversarial Networks**

In this paper, we propose Orthogonal Generative Adversarial Networks (O-GANs). We decompose the network of discriminator orthogonally and add an extra loss into the objective of common GANs, which can enforce discriminator become an effective encoder. The same extra loss can be embedded into any kind of GANs and there is almost no increase in computation. Furthermore, we discuss the principle of our method, which is relative to the fully-exploiting of the remaining degrees of freedom of discriminator. As we know, our solution is the simplest approach to train a generative adversarial network with auto-encoding ability.

**MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation**

Temporally locating and classifying action segments in long untrimmed videos is of particular interest to many applications like surveillance and robotics. While traditional approaches follow a two-step pipeline, by generating frame-wise probabilities and then feeding them to high-level temporal models, recent approaches use temporal convolutions to directly classify the video frames. In this paper, we introduce a multi-stage architecture for the temporal action segmentation task. Each stage features a set of dilated temporal convolutions to generate an initial prediction that is refined by the next one. This architecture is trained using a combination of a classification loss and a proposed smoothing loss that penalizes over-segmentation errors. Extensive evaluation shows the effectiveness of the proposed model in capturing long-range dependencies and recognizing action segments. Our model achieves state-of-the-art results on three challenging datasets: 50Salads, Georgia Tech Egocentric Activities (GTEA), and the Breakfast dataset.

**Statistical Guarantees for the Robustness of Bayesian Neural Networks**

We introduce a probabilistic robustness measure for Bayesian Neural Networks (BNNs), defined as the probability that, given a test point, there exists a point within a bounded set such that the BNN prediction differs between the two. Such a measure can be used, for instance, to quantify the probability of the existence of adversarial examples. Building on statistical verification techniques for probabilistic models, we develop a framework that allows us to estimate probabilistic robustness for a BNN with statistical guarantees, i.e., with a priori error and confidence bounds. We provide experimental comparison for several approximate BNN inference techniques on image classification tasks associated to MNIST and a two-class subset of the GTSRB dataset. Our results enable quantification of uncertainty of BNN predictions in adversarial settings.

### Like this:

Like Loading...

*Related*