Deep machine learning models are increasingly deployedin the wild for providing services to users. Adversaries maysteal the knowledge of these valuable models by trainingsubstitute models according to the inference results of thetargeted deployed models. Recent data-free model stealingmethods are shown effective to extract the knowledge of thetarget model without using real query examples, but they as-sume rich inference information, e.g., class probabilities andlogits. However, they are all based on competing generator-substitute networks and hence encounter training instability.In this paper we propose a data-free model stealing frame-work,MEGA, which is based on collaborative generator-substitute networks and only requires the target model toprovide label prediction for synthetic query examples. Thecore of our method is a model stealing optimization con-sisting of two collaborative models (i) the substitute modelwhich imitates the target model through the synthetic queryexamples and their inferred labels and (ii) the generatorwhich synthesizes images such that the confidence of thesubstitute model over each query example is maximized. Wepropose a novel coordinate descent training procedure andanalyze its convergence. We also empirically evaluate thetrained substitute model on three datasets and its applicationon black-box adversarial attacks. Our results show that theaccuracy of our trained substitute model and the adversarialattack success rate over it can be up to 33% and 40% higherthan state-of-the-art data-free black-box attacks.

In populous countries, pending legal cases have been growing exponentially. There is a need for developing techniques for processing and organizing legal documents. In this paper, we introduce a new corpus for structuring legal documents. In particular, we introduce a corpus of legal judgment documents in English that are segmented into topical and coherent parts. Each of these parts is annotated with a label coming from a list of pre-defined Rhetorical Roles. We develop baseline models for automatically predicting rhetorical roles in a legal document based on the annotated corpus. Further, we show the application of rhetorical roles to improve performance on the tasks of summarization and legal judgment prediction. We release the corpus and baseline model code along with the paper.

AI influences nearly 8 billion people and human & earth diverse ecosystems on an unprecedented scale. Startups accelerate to incorporate AI innovation as AI tools proliferate. UNESCO is the United Nations Educational, Scientific and Cultural Organization. The UNESCO recommendations on the ethics of AI recently adopted by member states provides a foundational global agreement on AI Ethics. The objectives ultimately drive emerging AI driven technologies that are trustworthy, safe, human-centered for the benefit of people and humanity.

Abstract--The COVID-19 (Coronavirus disease 2019) pandemic Index Terms--COVID-19, Deep learning, Early fusion, Feature has become a major global threat to human health and fusion, Uncertainty quantification. Such automatic systems are usually based on traditional machine learning or deep learning methods. We argue that the uncertainty of the model's predictions PCR has a low sensitivity. H.K. Lam is with the Centre for Robotics Research, Department of F. Karray is with the Centre for Pattern Analysis and Machine Intelligence, Department of Electrical and Computer Engineering, University of S. Hussain is with the System Administrator, Dibrugarh University, U. R. Acharya is with the Department of Electronics and Computer V. Makarenkov is with the Department of Computer Science, In recent years, deep learning models have had the Its areas of research and application have been growing widespread applicability not only in medical imaging field drastically. These models have allowed the information fusion to change from centralized also been extensively applied for COVID-19 detection. It is single node information fusion to distributed information critical to discriminate COVID-19 from other forms of pneumonia fusion. Farooq et al. [8] introduced an open-access Modern medicine nowadays depends on amalgamation dataset and the open-source code of their implementation of data and information from manifold sources that include using a CNN framework for distinguishing COVID-19 from structured imaging data, laboratory data, unstructured analogous pneumonia cohorts from chest X-ray images. The narrative data, and even observational or audio authors designed their COVIDResNet model by utilizing a data in some cases [22].

We introduce Monte-Carlo Attention (MCA), a randomized approximation method for reducing the computational cost of self-attention mechanisms in Transformer architectures. MCA exploits the fact that the importance of each token in an input sequence varies with respect to their attention scores; thus, some degree of error can be tolerable when encoding tokens with low attention. Using approximate matrix multiplication, MCA applies different error bounds to encode input tokens such that those with low attention scores are computed with relaxed precision, whereas errors of salient elements are minimized. MCA can operate in parallel with other attention optimization schemes and does not require model modification. We study the theoretical error bounds and demonstrate that MCA reduces attention complexity (in FLOPS) for various Transformer models by up to 11$\times$ in GLUE benchmarks without compromising model accuracy.

While artificial-intelligence-based methods suffer from lack of transparency, rule-based methods dominate in safety-critical systems. Yet, the latter cannot compete with the first ones in robustness to multiple requirements, for instance, simultaneously addressing safety, comfort, and efficiency. Hence, to benefit from both methods they must be joined in a single system. This paper proposes a decision making and control framework, which profits from advantages of both the rule- and machine-learning-based techniques while compensating for their disadvantages. The proposed method embodies two controllers operating in parallel, called Safety and Learned. A rule-based switching logic selects one of the actions transmitted from both controllers. The Safety controller is prioritized every time, when the Learned one does not meet the safety constraint, and also directly participates in the safe Learned controller training. Decision making and control in autonomous driving is chosen as the system case study, where an autonomous vehicle learns a multi-task policy to safely cross an unprotected intersection. Multiple requirements (i.e., safety, efficiency, and comfort) are set for vehicle operation. A numerical simulation is performed for the proposed framework validation, where its ability to satisfy the requirements and robustness to changing environment is successfully demonstrated.

Computing a Gaussian process (GP) posterior has a computational cost cubical in the number of historical points. A reformulation of the same GP posterior highlights that this complexity mainly depends on how many \emph{unique} historical points are considered. This can have important implication in active learning settings, where the set of historical points is constructed sequentially by the learner. We show that sequential black-box optimization based on GPs (GP-Opt) can be made efficient by sticking to a candidate solution for multiple evaluation steps and switch only when necessary. Limiting the number of switches also limits the number of unique points in the history of the GP. Thus, the efficient GP reformulation can be used to exactly and cheaply compute the posteriors required to run the GP-Opt algorithms. This approach is especially useful in real-world applications of GP-Opt with high switch costs (e.g. switching chemicals in wet labs, data/model loading in hyperparameter optimization). As examples of this meta-approach, we modify two well-established GP-Opt algorithms, GP-UCB and GP-EI, to switch candidates as infrequently as possible adapting rules from batched GP-Opt. These versions preserve all the theoretical no-regret guarantees while improving practical aspects of the algorithms such as runtime, memory complexity, and the ability of batching candidates and evaluating them in parallel.

This paper addresses the problem of precisely estimating the parameters of a stochastic model corresponding to branching processes. A branching process is a stochastic process consisting of collections of random variables indexed by the natural numbers. Branching processes are often used to describe population models Jagers (1989) and Athreya and Ney (2012); for example, models in the population genetics showing the genetic drift Burden and Simon (2016) Chen et al. (2017). In contrast to statistical approaches, branching processes enable the study of the dynamics of cell evolution and, as a consistence, have become a popular approach to cancer cell evolution research West et al., 2016. However, particularly in the case of cancer cell evolution, as well as in branching processes in general, the ultimate extinction of a population often occurs Devroye (1998). It is for this reason that with the initial uniform distribution of parameters, branching processes models tend to yield unevenly distributed data consisting of sparse and dense regions. The stochastic nature of the data is an another obstacle in estimating the parameters of a branching processes model, especially in the case of cancer cell evolution Nagornov et al. (2021). Moreover, simulations, based on a model of cell mutations, population evolution, and tumor/cancer subpopulations, commonly lead to the emergence of many clones and rarely to the appearance of cancer cells.

Although Transformer-based methods have significantly improved state-of-the-art results for long-term series forecasting, they are not only computationally expensive but more importantly, are unable to capture the global view of time series (e.g. overall trend). To address these problems, we propose to combine Transformer with the seasonal-trend decomposition method, in which the decomposition method captures the global profile of time series while Transformers capture more detailed structures. To further enhance the performance of Transformer for long-term prediction, we exploit the fact that most time series tend to have a sparse representation in well-known basis such as Fourier transform, and develop a frequency enhanced Transformer. Besides being more effective, the proposed method, termed as Frequency Enhanced Decomposed Transformer ({\bf FEDformer}), is more efficient than standard Transformer with a linear complexity to the sequence length. Our empirical studies with six benchmark datasets show that compared with state-of-the-art methods, FEDformer can reduce prediction error by $14.8\%$ and $22.6\%$ for multivariate and univariate time series, respectively. the code will be released soon.

The paper considers independent reinforcement learning (IRL) for multi-agent decision-making process in the paradigm of federated learning (FL). We show that FL can clearly improve the policy performance of IRL in terms of training efficiency and stability. However, since the policy parameters are trained locally and aggregated iteratively through a central server in FL, frequent information exchange incurs a large amount of communication overheads. To reach a good balance between improving the model's convergence performance and reducing the required communication and computation overheads, this paper proposes a system utility function and develops a consensus-based optimization scheme on top of the periodic averaging method, which introduces the consensus algorithm into FL for the exchange of a model's local gradients. This paper also provides novel convergence guarantees for the developed method, and demonstrates its superior effectiveness and efficiency in improving the system utility value through theoretical analyses and numerical simulation results.