亚洲男人的天堂2018av,欧美草比,久久久久久免费视频精选,国色天香在线看免费,久久久久亚洲av成人片仓井空

We study the asymptotic learning rates under linear and log-linear combination rules of belief vectors in a distributed hypothesis testing problem. We show that under both combination strategies, agents are able to learn the truth exponentially fast, with a faster rate under log-linear fusion. We examine the gap between the rates in terms of network connectivity and information diversity. We also provide closed-form expressions for special cases involving federated architectures and exchangeable networks.

Compression schemes have been extensively used in Federated Learning (FL) to reduce the communication cost of distributed learning. While most approaches rely on a bounded variance assumption of the noise produced by the compressor, this paper investigates the use of compression and aggregation schemes that produce a specific error distribution, e.g., Gaussian or Laplace, on the aggregated data. We present and analyze different aggregation schemes based on layered quantizers achieving exact error distribution. We provide different methods to leverage the proposed compression schemes to obtain compression-for-free in differential privacy applications. Our general compression methods can recover and improve standard FL schemes with Gaussian perturbations such as Langevin dynamics and randomized smoothing.

Quality assessment algorithms can be used to estimate the utility of a biometric sample for the purpose of biometric recognition. "Error versus Discard Characteristic" (EDC) plots, and "partial Area Under Curve" (pAUC) values of curves therein, are generally used by researchers to evaluate the predictive performance of such quality assessment algorithms. An EDC curve depends on an error type such as the "False Non Match Rate" (FNMR), a quality assessment algorithm, a biometric recognition system, a set of comparisons each corresponding to a biometric sample pair, and a comparison score threshold corresponding to a starting error. To compute an EDC curve, comparisons are progressively discarded based on the associated samples' lowest quality scores, and the error is computed for the remaining comparisons. Additionally, a discard fraction limit or range must be selected to compute pAUC values, which can then be used to quantitatively rank quality assessment algorithms. This paper discusses and analyses various details for this kind of quality assessment algorithm evaluation, including general EDC properties, interpretability improvements for pAUC values based on a hard lower error limit and a soft upper error limit, the use of relative instead of discrete rankings, stepwise vs. linear curve interpolation, and normalisation of quality scores to a [0, 100] integer range. We also analyse the stability of quantitative quality assessment algorithm rankings based on pAUC values across varying pAUC discard fraction limits and starting errors, concluding that higher pAUC discard fraction limits should be preferred. The analyses are conducted both with synthetic data and with real face image and fingerprint data, with a focus on general modality-independent conclusions for EDC evaluations. Various EDC alternatives are discussed as well.

To explain predictions made by complex machine learning models, many feature attribution methods have been developed that assign importance scores to input features. Some recent work challenges the robustness of these methods by showing that they are sensitive to input and model perturbations, while other work addresses this issue by proposing robust attribution methods. However, previous work on attribution robustness has focused primarily on gradient-based feature attributions, whereas the robustness of removal-based attribution methods is not currently well understood. To bridge this gap, we theoretically characterize the robustness properties of removal-based feature attributions. Specifically, we provide a unified analysis of such methods and derive upper bounds for the difference between intact and perturbed attributions, under settings of both input and model perturbations. Our empirical results on synthetic and real-world data validate our theoretical results and demonstrate their practical implications, including the ability to increase attribution robustness by improving the model's Lipschitz regularity.

The research on Reconfigurable Intelligent Surfaces (RISs) has dominantly been focused on physical-layer aspects and analyses of the achievable adaptation of the wireless propagation environment. Compared to that, questions related to system-level integration of RISs have received less attention. We address this research gap by analyzing the necessary control/signaling operations that are necessary to integrate RIS as a new type of wireless infrastructure element. We build a general model for evaluating the impact of control operations along two dimensions: i) the allocated bandwidth of the control channels (in-band and out-of-band), and ii) the rate selection for the data channel (multiplexing or diversity). Specifically, the second dimension results in two generic transmission schemes, one based on channel estimation and the subsequent optimization of the RIS, while the other is based on sweeping through predefined RIS phase configurations. We analyze the communication performance in multiple setups built along these two dimensions. While necessarily simplified, our analysis reveals the basic trade-offs in RIS-assisted communication and the associated control operations. The main contribution of the paper is a methodology for systematic evaluation of the control overhead in RIS-aided networks, regardless of the specific control schemes used.

A common explanation for the failure of out-of-distribution (OOD) generalization is that the model trained with empirical risk minimization (ERM) learns spurious features instead of invariant features. However, several recent studies challenged this explanation and found that deep networks may have already learned sufficiently good features for OOD generalization. Despite the contradictions at first glance, we theoretically show that ERM essentially learns both spurious and invariant features, while ERM tends to learn spurious features faster if the spurious correlation is stronger. Moreover, when fed the ERM learned features to the OOD objectives, the invariant feature learning quality significantly affects the final OOD performance, as OOD objectives rarely learn new features. Therefore, ERM feature learning can be a bottleneck to OOD generalization. To alleviate the reliance, we propose Feature Augmented Training (FeAT), to enforce the model to learn richer features ready for OOD generalization. FeAT iteratively augments the model to learn new features while retaining the already learned features. In each round, the retention and augmentation operations are performed on different subsets of the training data that capture distinct features. Extensive experiments show that FeAT effectively learns richer features thus boosting the performance of various OOD objectives.

In this work, our goal is to develop a theoretical framework that can eventually be used for analyzing the effectiveness of visual stories such as feature films to comic books. To develop this theoretical framework, we introduce a new story element called moments. Our conjecture is that any linear story such as the story of a feature film can be decomposed into a set of moments that follow each other. Moments are defined as the perception of the actions, interactions, and expressions of all characters or a single character during a given time period. We categorize the moments into two major types: story moments and discourse moments. Each type of moment can further be classified into three types, which we call universal storytelling moments. We believe these universal moments foster or deteriorate the emotional attachment of the audience to a particular character or the story. We present a methodology to catalog the occurrences of these universal moments as they are found in the story. The cataloged moments can be represented using curves or color strips. Therefore, we can visualize a character's journey through the story as either a 3D curve or a color strip. We also demonstrated that both story and discourse moments can be transformed into one lump-sum attraction parameter. The attraction parameter in time provides a function that can be plotted graphically onto a timeline illustrating changes in the emotional attachment of audience to a character or the story. By inspecting these functions the story analyst can analytically decipher the moments in the story where the attachment is being established, maintained, strengthened, or conversely where it is languishing.

We present an approach to the verification of systems for whose description some elements - constants or functions - are underspecified and can be regarded as parameters, and, in particular, describe a method for automatically generating constraints on such parameters under which certain safety conditions are guaranteed to hold. We present an implementation and illustrate its use on several examples.

The dominating NLP paradigm of training a strong neural predictor to perform one task on a specific dataset has led to state-of-the-art performance in a variety of applications (eg. sentiment classification, span-prediction based question answering or machine translation). However, it builds upon the assumption that the data distribution is stationary, ie. that the data is sampled from a fixed distribution both at training and test time. This way of training is inconsistent with how we as humans are able to learn from and operate within a constantly changing stream of information. Moreover, it is ill-adapted to real-world use cases where the data distribution is expected to shift over the course of a model's lifetime. The first goal of this thesis is to characterize the different forms this shift can take in the context of natural language processing, and propose benchmarks and evaluation metrics to measure its effect on current deep learning architectures. We then proceed to take steps to mitigate the effect of distributional shift on NLP models. To this end, we develop methods based on parametric reformulations of the distributionally robust optimization framework. Empirically, we demonstrate that these approaches yield more robust models as demonstrated on a selection of realistic problems. In the third and final part of this thesis, we explore ways of efficiently adapting existing models to new domains or tasks. Our contribution to this topic takes inspiration from information geometry to derive a new gradient update rule which alleviate catastrophic forgetting issues during adaptation.

Named entity recognition (NER) is the task to identify text spans that mention named entities, and to classify them into predefined categories such as person, location, organization etc. NER serves as the basis for a variety of natural language applications such as question answering, text summarization, and machine translation. Although early NER systems are successful in producing decent recognition accuracy, they often require much human effort in carefully designing rules or features. In recent years, deep learning, empowered by continuous real-valued vector representations and semantic composition through nonlinear processing, has been employed in NER systems, yielding stat-of-the-art performance. In this paper, we provide a comprehensive review on existing deep learning techniques for NER. We first introduce NER resources, including tagged NER corpora and off-the-shelf NER tools. Then, we systematically categorize existing works based on a taxonomy along three axes: distributed representations for input, context encoder, and tag decoder. Next, we survey the most representative methods for recent applied techniques of deep learning in new NER problem settings and applications. Finally, we present readers with the challenges faced by NER systems and outline future directions in this area.