在过去的几十年中，现代工业过程研究了几种具有成本效益的方法，以提高半导体制造的生产率和产量。虽然在促进实时监控和控制方面发挥重要作用，但行业中的数据驱动的软传感器在增强了晶圆故障诊断的深度学习方法时提供了竞争优势。尽管各个领域的深度学习方法取得了成功，但它们倾向于在多变化的软感测数据域上遭受不良性能。为了缓解这一点，我们提出了一种用于晶圆故障诊断分类任务的软感应集合器（卷积式变压器），主要由多头卷积模块组成，可获得快速和轻量级操作的卷曲的益处，以及能力通过多头设计相同的变压器来学习强大的表示。另一个关键问题是传统的学习范式倾向于在嘈杂和高度不平衡的软感测数据上遭受低性能。为了解决这个问题，我们使用基于课程的课程的损失函数增强了我们的软感测符合子模型，这有效地在培训的早期阶段和困难的阶段中学习易于样本。为了进一步展示我们拟议的架构的效用，我们对希捷技术的晶圆制造过程的各种工具进行了广泛的实验，这些工具与这项工作一起分享。据我们所知，这是第一次提出了课程，为软感测数据提出了基于课程的软感测符合子架构，我们的结果表明未来在软传感研究领域的使用中有很强的承诺。

在大数据的时代，基于数据驱动的分类已成为智能制造业的基本方法，以指导生产和优化检查。实践中获得的工业数据通常是由软传感器收集的时间序列数据，这是高度非线性，非间断，不平衡和嘈杂的。大多数现有的软传感机器学习模型侧重于捕获串联内部时间依赖关系或预定义的序列间相关性，同时忽略标签之间的相关性，每个实例同时与多个标签相关联。在本文中，我们提出了一种基于曲线的新颖的曲线图，用于多变量时间序列分类噪声和高度不平衡的软感测数据。所提出的基层能够在光谱域中捕获串联串联和串联系列依赖项; 2）通过叠加由统计共生信息构建的标签图来利用标签相关性; 3）从文本和数值域中使用注意机制学习功能; 4）利用未标记的数据并通过半监督学习缓解数据不平衡。与其他常用分类器的比较研究在希捷软感测数据上进行，实验结果验证了我们提出的方法的竞争性能。

随着近年来AI技术的快速发展，柔软传感区域的深层学习模型有很多研究。然而，该模型已经变得更加复杂，但数据集保持有限：研究人员用数百个数据样本拟合百万参数模型，这不足以行使其模型的有效性，因此在工业中实施时通常无法执行应用程序。为解决这一持久的问题，我们正在为公众提供大规模的高维时间序列制造传感器数据。我们展示了这些数据集上软感应变压器模型建模工业大数据的挑战和有效性。使用变压器，因为，它在自然语言处理中表现出优于最先进的技术，从直接应用到计算机视觉时也表现良好，而不引入图像特定的感应偏差。我们观察句子结构与传感器读数的相似性，并以自然语言中的句子类似的方式处理多变量传感器读数。高维时间序列数据被格式化成相同的嵌入式句子，并馈入变压器模型。结果表明，基于自动编码器和长短期存储器（LSTM）模型，变压器模型优于软感测场中的基准模型。据我们所知，我们是学术界或工业的第一支球队，以利用大规模数值软感测数据基准原始变压器模型的性能。

Large language models (LLMs) have shown impressive results across a variety of tasks while requiring little or no direct supervision. Further, there is mounting evidence that LLMs may have potential in information-seeking scenarios. We believe the ability of an LLM to attribute the text that it generates is likely to be crucial for both system developers and users in this setting. We propose and study Attributed QA as a key first step in the development of attributed LLMs. We develop a reproducable evaluation framework for the task, using human annotations as a gold standard and a correlated automatic metric that we show is suitable for development settings. We describe and benchmark a broad set of architectures for the task. Our contributions give some concrete answers to two key questions (How to measure attribution?, and How well do current state-of-the-art methods perform on attribution?), and give some hints as to how to address a third key question (How to build LLMs with attribution?).

This paper describes a prototype software and hardware platform to provide support to field operators during the inspection of surface defects of non-metallic pipes. Inspection is carried out by video filming defects created on the same surface in real-time using a "smart" helmet device and other mobile devices. The work focuses on the detection and recognition of the defects which appears as colored iridescence of reflected light caused by the diffraction effect arising from the presence of internal stresses in the inspected material. The platform allows you to carry out preliminary analysis directly on the device in offline mode, and, if a connection to the network is established, the received data is transmitted to the server for post-processing to extract information about possible defects that were not detected at the previous stage. The paper presents a description of the stages of design, formal description, and implementation details of the platform. It also provides descriptions of the models used to recognize defects and examples of the result of the work.

In this work, we demonstrate the offline FPGA realization of both recurrent and feedforward neural network (NN)-based equalizers for nonlinearity compensation in coherent optical transmission systems. First, we present a realization pipeline showing the conversion of the models from Python libraries to the FPGA chip synthesis and implementation. Then, we review the main alternatives for the hardware implementation of nonlinear activation functions. The main results are divided into three parts: a performance comparison, an analysis of how activation functions are implemented, and a report on the complexity of the hardware. The performance in Q-factor is presented for the cases of bidirectional long-short-term memory coupled with convolutional NN (biLSTM + CNN) equalizer, CNN equalizer, and standard 1-StpS digital back-propagation (DBP) for the simulation and experiment propagation of a single channel dual-polarization (SC-DP) 16QAM at 34 GBd along 17x70km of LEAF. The biLSTM+CNN equalizer provides a similar result to DBP and a 1.7 dB Q-factor gain compared with the chromatic dispersion compensation baseline in the experimental dataset. After that, we assess the Q-factor and the impact of hardware utilization when approximating the activation functions of NN using Taylor series, piecewise linear, and look-up table (LUT) approximations. We also show how to mitigate the approximation errors with extra training and provide some insights into possible gradient problems in the LUT approximation. Finally, to evaluate the complexity of hardware implementation to achieve 400G throughput, fixed-point NN-based equalizers with approximated activation functions are developed and implemented in an FPGA.

To circumvent the non-parallelizability of recurrent neural network-based equalizers, we propose knowledge distillation to recast the RNN into a parallelizable feedforward structure. The latter shows 38\% latency decrease, while impacting the Q-factor by only 0.5dB.

The paper discusses the improvement of the accuracy of an inertial navigation system created on the basis of MEMS sensors using machine learning (ML) methods. As input data for the classifier, we used infor-mation obtained from a developed laboratory setup with MEMS sensors on a sealed platform with the ability to adjust its tilt angles. To assess the effectiveness of the models, test curves were constructed with different values of the parameters of these models for each core in the case of a linear, polynomial radial basis function. The inverse regularization parameter was used as a parameter. The proposed algorithm based on MO has demonstrated its ability to correctly classify in the presence of noise typical for MEMS sensors, where good classification results were obtained when choosing the optimal values of hyperpa-rameters.

An algorithm and a program for detecting the boundaries of water bodies for the autopilot module of asurface robot are proposed. A method for detecting water objects on satellite maps by the method of finding a color in the HSV color space, using erosion, dilation - methods of digital image filtering is applied.The following operators for constructing contours on the image are investigated: the operators of Sobel,Roberts, Prewitt, and from them the one that detects the boundary more accurately is selected for thismodule. An algorithm for calculating the GPS coordinates of the contours is created. The proposed algorithm allows saving the result in a format suitable for the surface robot autopilot module.

The efficiency of using the YOLOV5 machine learning model for solving the problem of automatic de-tection and recognition of micro-objects in the marine environment is studied. Samples of microplankton and microplastics were prepared, according to which a database of classified images was collected for training an image recognition neural network. The results of experiments using a trained network to find micro-objects in photo and video images in real time are presented. Experimental studies have shown high efficiency, comparable to manual recognition, of the proposed model in solving problems of detect-ing micro-objects in the marine environment.