This study focuses on improving the optical character recognition (OCR) data for panels in the COMICS dataset, the largest dataset containing text and images from comic books. To do this, we developed a pipeline for OCR processing and labeling of comic books and created the first text detection and recognition datasets for western comics, called "COMICS Text+: Detection" and "COMICS Text+: Recognition". We evaluated the performance of state-of-the-art text detection and recognition models on these datasets and found significant improvement in word accuracy and normalized edit distance compared to the text in COMICS. We also created a new dataset called "COMICS Text+", which contains the extracted text from the textboxes in the COMICS dataset. Using the improved text data of COMICS Text+ in the comics processing model from resulted in state-of-the-art performance on cloze-style tasks without changing the model architecture. The COMICS Text+ dataset can be a valuable resource for researchers working on tasks including text detection, recognition, and high-level processing of comics, such as narrative understanding, character relations, and story generation. All the data and inference instructions can be accessed in https://github.com/gsoykan/comics_text_plus.

自我监督的学习允许AI系统使用不需要昂贵的标签的任务从大量数据中学习有效表示。模式崩溃，即为所有输入产生相同表示形式的模型，是许多自我监督学习方法的核心问题，可以使自我监督任务（例如匹配输入的变形变体）无效。在本文中，我们认为，同一输入的替代潜在表示之间信息最大化的直接应用自然解决了崩溃问题并实现了竞争性的经验结果。我们提出了一种自我监督的学习方法Corinfomax，该方法使用了基于二阶统计的共同信息度量，以反映其参数之间的相关性水平。在同一输入的替代表示之间最大化此相关信息度量有两个目的：（1）它通过生成具有非脱位协方差的特征向量来避免崩溃问题； （2）通过增加它们之间的线性依赖性，它在替代表示之间建立了相关性。提出的信息最大化客观的近似简化为基于欧几里得距离的目标函数，该目标函数由特征协方差矩阵的对数确定因素正规化。正则术语是针对特征空间退化的自然障碍。因此，除了避免完全输出崩溃到一个点外，提出的方法还通过鼓励信息在整个特征空间中的传播来防止尺寸崩溃。数值实验表明，相对于最先进的SSL方法，Corinfomax取得更好或竞争性的性能结果。

语言模型既展示了定量的改进，又展示了新的定性功能，随着规模的增加。尽管它们具有潜在的变革性影响，但这些新能力的特征却很差。为了为未来的研究提供信息，为破坏性的新模型能力做准备，并改善社会有害的效果，至关重要的是，我们必须了解目前和近乎未来的能力和语言模型的局限性。为了应对这一挑战，我们介绍了超越模仿游戏基准（Big Bench）。 Big Bench目前由204个任务组成，由132家机构的442位作者贡献。任务主题是多样的，从语言学，儿童发展，数学，常识性推理，生物学，物理学，社会偏见，软件开发等等。 Big-Bench专注于被认为超出当前语言模型的功能的任务。我们评估了OpenAI的GPT型号，Google内部密集变压器体系结构和大型基础上的开关稀疏变压器的行为，跨越了数百万到数十亿个参数。此外，一个人类专家评估者团队执行了所有任务，以提供强大的基准。研究结果包括：模型性能和校准都随规模改善，但绝对的术语（以及与评估者的性能相比）；在模型类中的性能非常相似，尽管带有稀疏性。逐渐和预测的任务通常涉及大量知识或记忆成分，而在临界规模上表现出“突破性”行为的任务通常涉及多个步骤或组成部分或脆性指标；社交偏见通常会随着含糊不清的环境而随着规模而增加，但这可以通过提示来改善。

如何最好地将涉及语言和视觉的多模式任务中的语言和感知处理整合在一起是一个重要的开放问题。在这项工作中，我们认为，以自上而下的方式使用语言的常见做法，直接通过高级视觉特征进行视觉关注，这可能不是最佳的。我们假设使用语言还可以调节从像素到高级功能的自下而上处理可以为整体性能带来好处。为了支持我们的主张，我们提出了一个基于U-NET的模型，并对两个语言视觉密集预测任务进行实验：引用表达式分割和语言引导的图像着色。我们比较结果，其中一个或自下而上的视觉分支都以语言为条件。我们的实验表明，除了自上而下的注意力外，使用语言来控制自下而上的视觉处理，从而为任务和实现竞争性表现提供更好的结果。我们的语言分析表明，自下而上的条件改善对象的细分，尤其是当输入文本是指低级视觉概念时。代码可在https://github.com/ilkerkesen/bvpr上找到。

Diffractive optical networks provide rich opportunities for visual computing tasks since the spatial information of a scene can be directly accessed by a diffractive processor without requiring any digital pre-processing steps. Here we present data class-specific transformations all-optically performed between the input and output fields-of-view (FOVs) of a diffractive network. The visual information of the objects is encoded into the amplitude (A), phase (P), or intensity (I) of the optical field at the input, which is all-optically processed by a data class-specific diffractive network. At the output, an image sensor-array directly measures the transformed patterns, all-optically encrypted using the transformation matrices pre-assigned to different data classes, i.e., a separate matrix for each data class. The original input images can be recovered by applying the correct decryption key (the inverse transformation) corresponding to the matching data class, while applying any other key will lead to loss of information. The class-specificity of these all-optical diffractive transformations creates opportunities where different keys can be distributed to different users; each user can only decode the acquired images of only one data class, serving multiple users in an all-optically encrypted manner. We numerically demonstrated all-optical class-specific transformations covering A-->A, I-->I, and P-->I transformations using various image datasets. We also experimentally validated the feasibility of this framework by fabricating a class-specific I-->I transformation diffractive network using two-photon polymerization and successfully tested it at 1550 nm wavelength. Data class-specific all-optical transformations provide a fast and energy-efficient method for image and data encryption, enhancing data security and privacy.

Recent advances in distributed artificial intelligence (AI) have led to tremendous breakthroughs in various communication services, from fault-tolerant factory automation to smart cities. When distributed learning is run over a set of wirelessly connected devices, random channel fluctuations and the incumbent services running on the same network impact the performance of both distributed learning and the coexisting service. In this paper, we investigate a mixed service scenario where distributed AI workflow and ultra-reliable low latency communication (URLLC) services run concurrently over a network. Consequently, we propose a risk sensitivity-based formulation for device selection to minimize the AI training delays during its convergence period while ensuring that the operational requirements of the URLLC service are met. To address this challenging coexistence problem, we transform it into a deep reinforcement learning problem and address it via a framework based on soft actor-critic algorithm. We evaluate our solution with a realistic and 3GPP-compliant simulator for factory automation use cases. Our simulation results confirm that our solution can significantly decrease the training delay of the distributed AI service while keeping the URLLC availability above its required threshold and close to the scenario where URLLC solely consumes all network resources.

Multispectral imaging has been used for numerous applications in e.g., environmental monitoring, aerospace, defense, and biomedicine. Here, we present a diffractive optical network-based multispectral imaging system trained using deep learning to create a virtual spectral filter array at the output image field-of-view. This diffractive multispectral imager performs spatially-coherent imaging over a large spectrum, and at the same time, routes a pre-determined set of spectral channels onto an array of pixels at the output plane, converting a monochrome focal plane array or image sensor into a multispectral imaging device without any spectral filters or image recovery algorithms. Furthermore, the spectral responsivity of this diffractive multispectral imager is not sensitive to input polarization states. Through numerical simulations, we present different diffractive network designs that achieve snapshot multispectral imaging with 4, 9 and 16 unique spectral bands within the visible spectrum, based on passive spatially-structured diffractive surfaces, with a compact design that axially spans ~72 times the mean wavelength of the spectral band of interest. Moreover, we experimentally demonstrate a diffractive multispectral imager based on a 3D-printed diffractive network that creates at its output image plane a spatially-repeating virtual spectral filter array with 2x2=4 unique bands at terahertz spectrum. Due to their compact form factor and computation-free, power-efficient and polarization-insensitive forward operation, diffractive multispectral imagers can be transformative for various imaging and sensing applications and be used at different parts of the electromagnetic spectrum where high-density and wide-area multispectral pixel arrays are not widely available.

Privacy-preserving inference via edge or encrypted computing paradigms encourages users of machine learning services to confidentially run a model on their personal data for a target task and only share the model's outputs with the service provider; e.g., to activate further services. Nevertheless, despite all confidentiality efforts, we show that a ''vicious'' service provider can approximately reconstruct its users' personal data by observing only the model's outputs, while keeping the target utility of the model very close to that of a ''honest'' service provider. We show the possibility of jointly training a target model (to be run at users' side) and an attack model for data reconstruction (to be secretly used at server's side). We introduce the ''reconstruction risk'': a new measure for assessing the quality of reconstructed data that better captures the privacy risk of such attacks. Experimental results on 6 benchmark datasets show that for low-complexity data types, or for tasks with larger number of classes, a user's personal data can be approximately reconstructed from the outputs of a single target inference task. We propose a potential defense mechanism that helps to distinguish vicious vs. honest classifiers at inference time. We conclude this paper by discussing current challenges and open directions for future studies. We open-source our code and results, as a benchmark for future work.

Federated learning (FL) is a promising approach to enable the future Internet of vehicles consisting of intelligent connected vehicles (ICVs) with powerful sensing, computing and communication capabilities. We consider a base station (BS) coordinating nearby ICVs to train a neural network in a collaborative yet distributed manner, in order to limit data traffic and privacy leakage. However, due to the mobility of vehicles, the connections between the BS and ICVs are short-lived, which affects the resource utilization of ICVs, and thus, the convergence speed of the training process. In this paper, we propose an accelerated FL-ICV framework, by optimizing the duration of each training round and the number of local iterations, for better convergence performance of FL. We propose a mobility-aware optimization algorithm called MOB-FL, which aims at maximizing the resource utilization of ICVs under short-lived wireless connections, so as to increase the convergence speed. Simulation results based on the beam selection and the trajectory prediction tasks verify the effectiveness of the proposed solution.

This technical report presents GPS++, the first-place solution to the Open Graph Benchmark Large-Scale Challenge (OGB-LSC 2022) for the PCQM4Mv2 molecular property prediction task. Our approach implements several key principles from the prior literature. At its core our GPS++ method is a hybrid MPNN/Transformer model that incorporates 3D atom positions and an auxiliary denoising task. The effectiveness of GPS++ is demonstrated by achieving 0.0719 mean absolute error on the independent test-challenge PCQM4Mv2 split. Thanks to Graphcore IPU acceleration, GPS++ scales to deep architectures (16 layers), training at 3 minutes per epoch, and large ensemble (112 models), completing the final predictions in 1 hour 32 minutes, well under the 4 hour inference budget allocated. Our implementation is publicly available at: https://github.com/graphcore/ogb-lsc-pcqm4mv2.