Animating portraits using speech has received growing attention in recent years, with various creative and practical use cases. An ideal generated video should have good lip sync with the audio, natural facial expressions and head motions, and high frame quality. In this work, we present SPACE, which uses speech and a single image to generate high-resolution, and expressive videos with realistic head pose, without requiring a driving video. It uses a multi-stage approach, combining the controllability of facial landmarks with the high-quality synthesis power of a pretrained face generator. SPACE also allows for the control of emotions and their intensities. Our method outperforms prior methods in objective metrics for image quality and facial motions and is strongly preferred by users in pair-wise comparisons. The project website is available at https://deepimagination.cc/SPACE/

我们介绍了ADAVIT，一种可自适应地调整视觉变压器（VIT）推理成本的方法，用于不同复杂性的图像。 Adavit通过自动减少在网络中处理的视觉变压器中的令牌数量作为推理进行的令牌的数量来实现这一目标。我们为此任务进行重新格式化自适应计算时间（ACT），扩展为丢弃冗余空间令牌。视觉变换器的吸引力架构属性使我们的自适应令牌减少机制能够加速推理而不修改网络架构或推理硬件。我们展示了ADAVIT不需要额外的参数或子网来停止，因为我们基于自适应停止在原始网络参数上的学习。我们进一步引入了与现有行为方法相比稳定培训的分布先前正则化。在图像分类任务（ImageNet1K）上，我们表明我们提出的Adavit在过滤信息丰富的空间特征和削减整体计算上产生了高效率。所提出的方法将Deit-Tiny的吞吐量提高了62％并除去了38％，只有0.3％的精度下降，优于大边距。

现有条件图像综合框架基于单个模态中的用户输入生成图像，例如文本，分段，草图或样式参考。它们通常无法在可用时利用多式联运用户输入，这降低了它们的实用性。为了解决这一限制，我们提出了专家级别的生成对抗性网络（PoE-GaN）框架，其可以合成在多个输入模态或其任何子集上调节的图像，即使是空集。 Poe-GaN包括专家级别的发电机和多模式多尺寸投影鉴别器。通过我们精心设计的培训计划，Poe-GaN学习以高质量和多样性合成图像。除了在多模式条件图像合成中推进现有技术之外，PoE-GaN还优于在单向设置中测试时最佳现有的单峰条件图像合成方法。该项目网站在https://deepimagination.github.io/poe-gan提供。

Structural pruning of neural network parameters reduces computation, energy, and memory transfer costs during inference. We propose a novel method that estimates the contribution of a neuron (filter) to the final loss and iteratively removes those with smaller scores. We describe two variations of our method using the first and secondorder Taylor expansions to approximate a filter's contribution. Both methods scale consistently across any network layer without requiring per-layer sensitivity analysis and can be applied to any kind of layer, including skip connections. For modern networks trained on ImageNet, we measured experimentally a high (>93%) correlation between the contribution computed by our methods and a reliable estimate of the true importance. Pruning with the proposed methods leads to an improvement over state-ofthe-art in terms of accuracy, FLOPs, and parameter reduction. On ResNet-101, we achieve a 40% FLOPS reduction by removing 30% of the parameters, with a loss of 0.02% in the top-1 accuracy on ImageNet. Code is available at https://github.com/NVlabs/Taylor_pruning.

This paper presents a method for adding multiple tasks to a single deep neural network while avoiding catastrophic forgetting. Inspired by network pruning techniques, we exploit redundancies in large deep networks to free up parameters that can then be employed to learn new tasks. By performing iterative pruning and network re-training, we are able to sequentially "pack" multiple tasks into a single network while ensuring minimal drop in performance and minimal storage overhead. Unlike prior work that uses proxy losses to maintain accuracy on older tasks, we always optimize for the task at hand. We perform extensive experiments on a variety of network architectures and largescale datasets, and observe much better robustness against catastrophic forgetting than prior work. In particular, we are able to add three fine-grained classification tasks to a single ImageNet-trained VGG-16 network and achieve accuracies close to those of separately trained networks for each task. Code available at https://github.com/ arunmallya/packnet

We introduce a new tool for stochastic convex optimization (SCO): a Reweighted Stochastic Query (ReSQue) estimator for the gradient of a function convolved with a (Gaussian) probability density. Combining ReSQue with recent advances in ball oracle acceleration [CJJJLST20, ACJJS21], we develop algorithms achieving state-of-the-art complexities for SCO in parallel and private settings. For a SCO objective constrained to the unit ball in $\mathbb{R}^d$, we obtain the following results (up to polylogarithmic factors). We give a parallel algorithm obtaining optimization error $\epsilon_{\text{opt}}$ with $d^{1/3}\epsilon_{\text{opt}}^{-2/3}$ gradient oracle query depth and $d^{1/3}\epsilon_{\text{opt}}^{-2/3} + \epsilon_{\text{opt}}^{-2}$ gradient queries in total, assuming access to a bounded-variance stochastic gradient estimator. For $\epsilon_{\text{opt}} \in [d^{-1}, d^{-1/4}]$, our algorithm matches the state-of-the-art oracle depth of [BJLLS19] while maintaining the optimal total work of stochastic gradient descent. We give an $(\epsilon_{\text{dp}}, \delta)$-differentially private algorithm which, given $n$ samples of Lipschitz loss functions, obtains near-optimal optimization error and makes $\min(n, n^2\epsilon_{\text{dp}}^2 d^{-1}) + \min(n^{4/3}\epsilon_{\text{dp}}^{1/3}, (nd)^{2/3}\epsilon_{\text{dp}}^{-1})$ queries to the gradients of these functions. In the regime $d \le n \epsilon_{\text{dp}}^{2}$, where privacy comes at no cost in terms of the optimal loss up to constants, our algorithm uses $n + (nd)^{2/3}\epsilon_{\text{dp}}^{-1}$ queries and improves recent advancements of [KLL21, AFKT21]. In the moderately low-dimensional setting $d \le \sqrt n \epsilon_{\text{dp}}^{3/2}$, our query complexity is near-linear.

Periocular refers to the region of the face that surrounds the eye socket. This is a feature-rich area that can be used by itself to determine the identity of an individual. It is especially useful when the iris or the face cannot be reliably acquired. This can be the case of unconstrained or uncooperative scenarios, where the face may appear partially occluded, or the subject-to-camera distance may be high. However, it has received revived attention during the pandemic due to masked faces, leaving the ocular region as the only visible facial area, even in controlled scenarios. This paper discusses the state-of-the-art of periocular biometrics, giving an overall framework of its most significant research aspects.

Recently, Smart Video Surveillance (SVS) systems have been receiving more attention among scholars and developers as a substitute for the current passive surveillance systems. These systems are used to make the policing and monitoring systems more efficient and improve public safety. However, the nature of these systems in monitoring the public's daily activities brings different ethical challenges. There are different approaches for addressing privacy issues in implementing the SVS. In this paper, we are focusing on the role of design considering ethical and privacy challenges in SVS. Reviewing four policy protection regulations that generate an overview of best practices for privacy protection, we argue that ethical and privacy concerns could be addressed through four lenses: algorithm, system, model, and data. As an case study, we describe our proposed system and illustrate how our system can create a baseline for designing a privacy perseverance system to deliver safety to society. We used several Artificial Intelligence algorithms, such as object detection, single and multi camera re-identification, action recognition, and anomaly detection, to provide a basic functional system. We also use cloud-native services to implement a smartphone application in order to deliver the outputs to the end users.

As text generated by large language models proliferates, it becomes vital to understand how humans engage with such text, and whether or not they are able to detect when the text they are reading did not originate with a human writer. Prior work on human detection of generated text focuses on the case where an entire passage is either human-written or machine-generated. In this paper, we study a more realistic setting where text begins as human-written and transitions to being generated by state-of-the-art neural language models. We show that, while annotators often struggle at this task, there is substantial variance in annotator skill and that given proper incentives, annotators can improve at this task over time. Furthermore, we conduct a detailed comparison study and analyze how a variety of variables (model size, decoding strategy, fine-tuning, prompt genre, etc.) affect human detection performance. Finally, we collect error annotations from our participants and use them to show that certain textual genres influence models to make different types of errors and that certain sentence-level features correlate highly with annotator selection. We release the RoFT dataset: a collection of over 21,000 human annotations paired with error classifications to encourage future work in human detection and evaluation of generated text.