在本文中，我们询问视觉变形金刚（VIT）是否可以作为改善机器学习模型对抗逃避攻击的对抗性鲁棒性的基础结构。尽管较早的作品集中在改善卷积神经网络上，但我们表明VIT也非常适合对抗训练以实现竞争性能。我们使用自定义的对抗训练配方实现了这一目标，该配方是在Imagenet数据集的一部分上使用严格的消融研究发现的。与卷积相比，VIT的规范培训配方建议强大的数据增强，部分是为了补偿注意力模块的视力归纳偏置。我们表明，该食谱在用于对抗训练时可实现次优性能。相比之下，我们发现省略所有重型数据增强，并添加一些额外的零件（$ \ varepsilon $ -Warmup和更大的重量衰减），从而大大提高了健壮的Vits的性能。我们表明，我们的配方在完整的Imagenet-1k上概括了不同类别的VIT体系结构和大规模模型。此外，调查了模型鲁棒性的原因，我们表明，在使用我们的食谱时，在训练过程中产生强烈的攻击更加容易，这会在测试时提高鲁棒性。最后，我们通过提出一种量化对抗性扰动的语义性质并强调其与模型的鲁棒性的相关性来进一步研究对抗训练的结果。总体而言，我们建议社区应避免将VIT的规范培训食谱转换为在对抗培训的背景下进行强大的培训和重新思考常见的培训选择。

作为研究界，我们仍然缺乏对对抗性稳健性的进展的系统理解，这通常使得难以识别训练强大模型中最有前途的想法。基准稳健性的关键挑战是，其评估往往是出错的导致鲁棒性高估。我们的目标是建立对抗性稳健性的标准化基准，尽可能准确地反映出考虑在合理的计算预算范围内所考虑的模型的稳健性。为此，我们首先考虑图像分类任务并在允许的型号上引入限制（可能在将来宽松）。我们评估了与AutoAtrack的对抗鲁棒性，白和黑箱攻击的集合，最近在大规模研究中显示，与原始出版物相比，改善了几乎所有稳健性评估。为防止对自动攻击进行新防御的过度适应，我们欢迎基于自适应攻击的外部评估，特别是在自动攻击稳健性潜在高估的地方。我们的排行榜，托管在https://robustbench.github.io/，包含120多个模型的评估，并旨在反映在$ \ ell_ \ infty $的一套明确的任务上的图像分类中的当前状态 - 和$ \ ell_2 $ -Threat模型和共同腐败，未来可能的扩展。此外，我们开源源是图书馆https://github.com/robustbench/robustbench，可以提供对80多个强大模型的统一访问，以方便他们的下游应用程序。最后，根据收集的模型，我们分析了稳健性对分布换档，校准，分配检测，公平性，隐私泄漏，平滑度和可转移性的影响。

Evaluating and comparing text-to-image models is a challenging problem. Significant advances in the field have recently been made, piquing interest of various industrial sectors. As a consequence, a gold standard in the field should cover a variety of tasks and application contexts. In this paper a novel evaluation approach is experimented, on the basis of: (i) a curated data set, made by high-quality royalty-free image-text pairs, divided into ten categories; (ii) a quantitative metric, the CLIP-score, (iii) a human evaluation task to distinguish, for a given text, the real and the generated images. The proposed method has been applied to the most recent models, i.e., DALLE2, Latent Diffusion, Stable Diffusion, GLIDE and Craiyon. Early experimental results show that the accuracy of the human judgement is fully coherent with the CLIP-score. The dataset has been made available to the public.

One of the common traits of past and present approaches for Semantic Role Labeling (SRL) is that they rely upon discrete labels drawn from a predefined linguistic inventory to classify predicate senses and their arguments. However, we argue this need not be the case. In this paper, we present an approach that leverages Definition Modeling to introduce a generalized formulation of SRL as the task of describing predicate-argument structures using natural language definitions instead of discrete labels. Our novel formulation takes a first step towards placing interpretability and flexibility foremost, and yet our experiments and analyses on PropBank-style and FrameNet-style, dependency-based and span-based SRL also demonstrate that a flexible model with an interpretable output does not necessarily come at the expense of performance. We release our software for research purposes at https://github.com/SapienzaNLP/dsrl.

Image generation and image completion are rapidly evolving fields, thanks to machine learning algorithms that are able to realistically replace missing pixels. However, generating large high resolution images, with a large level of details, presents important computational challenges. In this work, we formulate the image generation task as completion of an image where one out of three corners is missing. We then extend this approach to iteratively build larger images with the same level of detail. Our goal is to obtain a scalable methodology to generate high resolution samples typically found in satellite imagery data sets. We introduce a conditional progressive Generative Adversarial Networks (GAN), that generates the missing tile in an image, using as input three initial adjacent tiles encoded in a latent vector by a Wasserstein auto-encoder. We focus on a set of images used by the United Nations Satellite Centre (UNOSAT) to train flood detection tools, and validate the quality of synthetic images in a realistic setup.

Parameter-efficient fine-tuning (PEFT) methods can adapt large language models to downstream tasks by training a small amount of newly added parameters. In multi-task settings, PEFT adapters typically train on each task independently, inhibiting transfer across tasks, or on the concatenation of all tasks, which can lead to negative interference. To address this, Polytropon (Ponti et al.) jointly learns an inventory of PEFT adapters and a routing function to share variable-size sets of adapters across tasks. Subsequently, adapters can be re-combined and fine-tuned on novel tasks even with limited data. In this paper, we investigate to what extent the ability to control which adapters are active for each task leads to sample-efficient generalization. Thus, we propose less expressive variants where we perform weighted averaging of the adapters before few-shot adaptation (Poly-mu) instead of learning a routing function. Moreover, we introduce more expressive variants where finer-grained task-adapter allocation is learned through a multi-head routing function (Poly-S). We test these variants on three separate benchmarks for multi-task learning. We find that Poly-S achieves gains on all three (up to 5.3 points on average) over strong baselines, while incurring a negligible additional cost in parameter count. In particular, we find that instruction tuning, where models are fully fine-tuned on natural language instructions for each task, is inferior to modular methods such as Polytropon and our proposed variants.

With climate change predicted to increase the likelihood of landslide events, there is a growing need for rapid landslide detection technologies that help inform emergency responses. Synthetic Aperture Radar (SAR) is a remote sensing technique that can provide measurements of affected areas independent of weather or lighting conditions. Usage of SAR, however, is hindered by domain knowledge that is necessary for the pre-processing steps and its interpretation requires expert knowledge. We provide simplified, pre-processed, machine-learning ready SAR datacubes for four globally located landslide events obtained from several Sentinel-1 satellite passes before and after a landslide triggering event together with segmentation maps of the landslides. From this dataset, using the Hokkaido, Japan datacube, we study the feasibility of SAR-based landslide detection with supervised deep learning (DL). Our results demonstrate that DL models can be used to detect landslides from SAR data, achieving an Area under the Precision-Recall curve exceeding 0.7. We find that additional satellite visits enhance detection performance, but that early detection is possible when SAR data is combined with terrain information from a digital elevation model. This can be especially useful for time-critical emergency interventions. Code is made publicly available at https://github.com/iprapas/landslide-sar-unet.

现有的数据驱动方法用于披上姿势的人体，尽管有效，但无法处理任意拓扑的服装，并且通常不是端到端的。为了解决这些局限性，我们提出了一条端到端可区分管道，该管道用隐式表面表示服装，并学习以铰接式身体模型的形状和姿势参数为条件的皮肤场。为了限制身体的插入和人工制品，我们提出了一种解释意识的训练数据的预处理策略和新颖的训练损失，在覆盖服装的同时惩罚了自身交流。我们证明，我们的方法可以针对最新方法产生更准确的结果和变形。此外，我们表明我们的方法凭借其端到端的可不同性，可以从图像观察中共同恢复身体和服装参数，这是以前的工作无法做到的。

由于技术成本的降低和卫星发射的增加，卫星图像变得越来越流行和更容易获得。除了提供仁慈的目的外，还可以出于恶意原因（例如错误信息）使用卫星数据。事实上，可以依靠一般图像编辑工具来轻松操纵卫星图像。此外，随着深层神经网络（DNN）的激增，可以生成属于各种领域的现实合成图像，与合成生成的卫星图像的扩散有关的其他威胁正在出现。在本文中，我们回顾了关于卫星图像的产生和操纵的最新技术（SOTA）。特别是，我们既关注从头开始的合成卫星图像的产生，又要通过图像转移技术对卫星图像进行语义操纵，包括从一种类型的传感器到另一种传感器获得的图像的转换。我们还描述了迄今已研究的法医检测技术，以对合成图像伪造进行分类和检测。虽然我们主要集中在法医技术上明确定制的，该技术是针对AI生成的合成内容物的检测，但我们还审查了一些用于一般剪接检测的方法，这些方法原则上也可以用于发现AI操纵图像

自主赛车是一项研究领域，由于它将自动驾驶算法推向极限，并作为一般自主驾驶的催化剂。对于规模的自主赛车平台，计算约束和复杂性通常会限制模型预测控制（MPC）的使用。结果，几何控制器是最常部署的控制器。它们在实施和操作简单性的同时被证明是性能。然而，他们固有地缺乏模型动力学的结合，因此将赛车限制在可以忽略轮胎滑动的速度域。本文介绍了基于模型和加速度的追求（MAP）基于高性能模型的轨迹跟踪算法，该算法在利用轮胎动力学的同时保留了几何方法的简单性。与最先进的几何控制器相比，所提出的算法允许在前所未有的速度上准确跟踪轨迹。在横向跟踪误差方面，在实验上验证了地图控制器，并胜过参考几何控制器四倍，以高达11m/s的测试速度产生0.055m的跟踪误差。