Object detection with 3D radar is essential for 360○automotive perception, but radar’s long wavelengths produce sparse and irregular reflections that challenge traditional grid and sequence-based convolutional and transformer detectors. This paper introduces Graph Query Networks (GQN), an attention-based framework that models objects sensed by radar as graphs, to extract individualized relational and contextual features. GQN employs a novel concept of graph queries to dynamically attend over the bird’s-eye view (BEV) space, constructing object-specific graphs processed by two novel modules: EdgeFocus for relational reasoning and DeepContext Pooling for contextual aggregation. On the NuScenes dataset, GQN improves relative mAP by up to +53%, including a +8.2% gain over the strongest prior radar method, while reducing peak graph construction overhead by 80% with moderate FLOPs cost.

Modern manufacturing relies heavily on fusion welding processes, including gas metal arc welding (GMAW). Despite significant advances in machine learning-based quality prediction, current models exhibit critical limitations when confronted with the inherent distribution shifts that occur in dynamic manufacturing environments. In this work, we extend the VQ-VAE Transformer architecture-previously demonstrating state-of-the-art performance in weld quality prediction-by leveraging its autoregressive loss as a reliable out-of-distribution (OOD) detection mechanism. Our approach exhibits superior performance compared to conventional reconstruction methods, embedding error-based techniques, and other established baselines. By integrating OOD detection with continual learning strategies, we optimize model adaptation, triggering updates only when necessary and thereby minimizing costly labeling requirements. We introduce a novel quantitative metric that simultaneously evaluates OOD detection capability while interpreting in-distribution performance. Experimental validation in real-world welding scenarios demonstrates that our framework effectively maintains robust quality prediction capabilities across significant distribution shifts, addressing critical challenges in dynamic manufacturing environments where process parameters frequently change. This research makes a substantial contribution to applied artificial intelligence by providing an explainable and at the same time adaptive solution for quality assurance in dynamic manufacturing processes-a crucial step towards robust, practical AI systems in the industrial environment.

Abstract The field of emergent language represents a novel area of research within the domain of artificial intelligence, particularly within the context of multi-agent reinforcement learning. Although the concept of studying language emergence is not new, early approaches were primarily concerned with explaining human language formation, with little consideration given to its potential utility for artificial agents. In contrast, studies based on reinforcement learning aim to develop communicative capabilities in agents that are comparable to or even superior to human language. Thus, they extend beyond the learned statistical representations that are common in natural language processing research. This gives rise to a number of fundamental questions, from the prerequisites for language emergence to the criteria for measuring its success. This paper addresses these questions by providing a comprehensive review of relevant scientific publications on emergent language in artificial intelligence. Its objective is to serve as a reference for researchers interested in or proficient in the field. Consequently, the main contributions are the definition and overview of the prevailing terminology, the analysis of existing evaluation methods and metrics, and the description of the identified research gaps.

Modern manufacturing relies heavily on fusion welding processes, including gas metal arc welding (GMAW), which efficiently converts electrical energy into thermal energy to join metals. Despite decades of research and extensive application in the automotive and aerospace sectors, weld quality assessment in the GMAW process remains a major challenge. This paper presents a novel learning-based approach relying on a vector quantised variational autoencoder (VQ-VAE) for data representation. In addition, we are the first to provide a time series dataset to the research community that combines labeled and unlabeled time series data from the GMAW domain, thereby enabling further research. The core idea of our approach consists of two stages: In the first stage, we use a learned automatic extraction of local features of the input signal using a VQ-VAE architecture. Based on this, in the second stage, we use a transformer model that processes the discretized features and performs weld quality prediction and classification. Our approach addresses real-world scenarios and improves the prediction of quality and fill existing data gaps by providing a reliable approach for quality assessment during manufacturing based on sensor data.

The digitization of manufacturing processes opens up the possibility of using machine learning methods on process data to predict future product quality. Based on the model predictions, quality improvement actions can be taken at an early stage. However, significant challenges must be overcome to successfully implement the predictions. Production lines are subject to hardware and memory limitations and are characterized by constant changes in quality influencing factors. In this paper, we address these challenges and present an online prediction approach for real-world manufacturing processes. On the one hand, it includes methods for feature extraction and selection from multimodal process and sensor data. On the other hand, a continual learning method based on memory-aware synapses is developed to efficiently train an artificial neural network over process changes. We deploy and evaluate the approach in a windshield production process. Our experimental evaluation shows that the model can accurately predict windshield quality and achieve significant process improvement. By comparing with other learning strategies such as transfer learning, we also show that the continual learning method both prevents catastrophic forgetting of the model and maintains its data efficiency.

The assembly line production of electrical consumer products is a highly streamlined process in which the product quality is continuously evaluated using automated checks. However, some products include manual processing due to customer requests that are not covered by standardized production plans. In such situations, quality issues frequently remain unnoticed leading to high reversal costs and customer dissatisfaction. We address this problem in a practical case study for a specific product family that is subject to highly versatile and error-prone configurations of externally exposed hardware connectors. In this setting, the worker must be visually assisted such that potentially faulty configurations are highlighted. Therefore, we investigate the applicability of state-of-the-art approaches for Anomaly Detection (AD) and Anomaly Localization (AL) on image data using pre-trained models and normalizing flows and compare against baseline Variational Auto-Encoders (VAEs). We show that those methods are not only applicable to well-established benchmarks on industrial image data but also have the potential to be used in a practical use case.

Recently, time series classification methods based on Convolutional Neural Networks (CNNs) have demonstrated state-of-the-art performance outperforming former ensemble-based methods like HIVE-COTE on a multitude of time series datasets. Inspired by the current rise of Deep Neural Networks (DNNs) end-to-end classifiers for time series classification, we propose utilisation of Deformable Convolutional Neural Networks (Deformable CNNs), which have already proven to drastically enhance classification performance on image classification tasks. Our aim is to evaluate the applicability of such methods on the practical use-case of a German railway provider, in which sensory data from railway switches is employed to detect and classify faults in switching operation. Prior to any classification, we have to address two main issues, which is that the available data is in a raw, unlabelled format and the contained time series have vastly varying length. We cope by applying extensive pre-processing and semi-supervised labelling. As baseline classifier, we use a conventional KNN classifier that is tailored to enable handling of sensory data. Finally, we compare the baseline classifier against more advanced DNN classifiers and discuss their feasibility in general and in context of our use-case.

In the field of production, machine learning offers great potentials to develop innovative solutions for optimization or automation. However, it faces challenges with regard to the availability of data and the high training effort of the learning models in the event of changes in the production process. In this paper, we address these challenges by introducing deep transfer learning for production. We demonstrate its potentials and benefits in a real application for predictive quality in injection molding and propose a novel approach for the continual training of neural networks across manufactured products. By creating a neural network that leverages knowledge from previous products without forgetting them, the approach shows better learning rates and more accurate predictions while requiring much less data for training. Our code is publicly available1to reproduce our results and build upon them.