Michael Friedrich

Zhining Hu, Tobias Schlosser, Michael Friedrich et al.

In semiconductor manufacturing, the wafer dicing process is central yet vulnerable to defects that significantly impair yield - the proportion of defect-free chips. Deep neural networks are the current state of the art in (semi-)automated visual inspection. However, they are notoriously known to require a particularly large amount of data for model training. To address these challenges, we explore the application of generative adversarial networks (GAN) for image data augmentation and classification of semiconductor wafer dicing induced defects to enhance the variety and balance of training data for visual inspection systems. With this approach, synthetic yet realistic images are generated that mimic real-world dicing defects. We employ three different GAN variants for high-resolution image synthesis: Deep Convolutional GAN (DCGAN), CycleGAN, and StyleGAN3. Our work-in-progress results demonstrate that improved classification accuracies can be obtained, showing an average improvement of up to 23.1 % from 65.1 % (baseline experiment) to 88.2 % (DCGAN experiment) in balanced accuracy, which may enable yield optimization in production.

Karthik Suresh, Tushar Vatsa, Tracy King et al.

Late-interaction retrieval models rely on hard maximum similarity (MaxSim) to aggregate token-level similarities. Although effective, this winner-take-all pooling rule may structurally bias training dynamics. We provide a mechanistic study of gradient routing and robustness in MaxSim-based retrieval. In a controlled synthetic environment with in-batch contrastive training, we demonstrate that MaxSim induces significantly higher patch-level gradient concentration than smoother alternatives such as Top-k pooling and softmax aggregation. While sparse routing can improve early discrimination, it also increases sensitivity to document length: as the number of document patches grows, MaxSim degrades more sharply than mild smoothing variants. We corroborate these findings on a real-world multi-vector retrieval benchmark, where controlled document-length sweeps reveal similar brittleness under hard max pooling. Together, our results isolate pooling-induced gradient concentration as a structural property of late-interaction retrieval and highlight a sparsity-robustness tradeoff. These findings motivate principled alternatives to hard max pooling in multi-vector retrieval systems.

Dewang Sultania, Zhaoyu Lu, Twisha Naik et al.

Domain specific question answering is an evolving field that requires specialized solutions to address unique challenges. In this paper, we show that a hybrid approach combining a fine-tuned dense retriever with keyword based sparse search methods significantly enhances performance. Our system leverages a linear combination of relevance signals, including cosine similarity from dense retrieval, BM25 scores, and URL host matching, each with tunable boost parameters. Experimental results indicate that this hybrid method outperforms our single-retriever system, achieving improved accuracy while maintaining robust contextual grounding. These findings suggest that integrating multiple retrieval methodologies with weighted scoring effectively addresses the complexities of domain specific question answering in enterprise settings.

Vibha Belavadi, Tushar Vatsa, Dewang Sultania et al. · deepmind

This paper addresses fine-tuning Large Language Models (LLMs) for function calling tasks when real user interaction data is unavailable. In digital content creation tools, where users express their needs through natural language queries that must be mapped to API calls, the lack of real-world task-specific data and privacy constraints for training on it necessitate synthetic data generation. Existing approaches to synthetic data generation fall short in diversity and complexity, failing to replicate real-world data distributions and leading to suboptimal performance after LLM fine-tuning. We present a novel router-based architecture that leverages domain resources like content metadata and structured knowledge graphs, along with text-to-text and vision-to-text language models to generate high-quality synthetic training data. Our architecture's flexible routing mechanism enables synthetic data generation that matches observed real-world distributions, addressing a fundamental limitation of traditional approaches. Evaluation on a comprehensive set of real user queries demonstrates significant improvements in both function classification accuracy and API parameter selection. Models fine-tuned with our synthetic data consistently outperform traditional approaches, establishing new benchmarks for function calling tasks.

Frederik Beuth, Tobias Schlosser, Michael Friedrich et al.

It is a long-term goal to transfer biological processing principles as well as the power of human recognition into machine vision and engineering systems. One of such principles is visual attention, a smart human concept which focuses processing on a part of a scene. In this contribution, we utilize attention to improve the automatic detection of defect patterns for wafers within the domain of semiconductor manufacturing. Previous works in the domain have often utilized classical machine learning approaches such as KNNs, SVMs, or MLPs, while a few have already used modern approaches like deep neural networks (DNNs). However, one problem in the domain is that the faults are often very small and have to be detected within a larger size of the chip or even the wafer. Therefore, small structures in the size of pixels have to be detected in a vast amount of image data. One interesting principle of the human brain for solving this problem is visual attention. Hence, we employ here a biologically plausible model of visual attention for automatic visual inspection. We propose a hybrid system of visual attention and a deep neural network. As demonstrated, our system achieves among other decisive advantages an improvement in accuracy from 81% to 92%, and an increase in accuracy for detecting faults from 67% to 88%. Hence, the error rates are reduced from 19% to 8%, and notably from 33% to 12% for detecting a fault in a chip. These results show that attention can greatly improve the performance of visual inspection systems. Furthermore, we conduct a broad evaluation, identifying specific advantages of the biological attention model in this application, and benchmarks standard deep learning approaches as an alternative with and without attention. This work is an extended arXiv version of the original conference article published in "IECON 2020", which has been extended regarding visual attention.

Tobias Schlosser, Michael Friedrich, Danny Kowerko

Inspired by the human visual perception system, hexagonal image processing in the context of machine learning deals with the development of image processing systems that combine the advantages of evolutionary motivated structures based on biological models. While conventional state-of-the-art image processing systems of recording and output devices almost exclusively utilize square arranged methods, their hexagonal counterparts offer a number of key advantages that can benefit both researchers and users. This contribution serves as a general application-oriented approach the synthesis of the therefore designed hexagonal image processing framework, called Hexnet, the processing steps of hexagonal image transformation, and dependent methods. The results of our created test environment show that the realized framework surpasses current approaches of hexagonal image processing systems, while hexagonal artificial neural networks can benefit from the implemented hexagonal architecture. As hexagonal lattice format based deep neural networks, also called H-DNN, can be compared to their square counterparts by transforming classical square lattice based data sets into their hexagonal representation, they can also result in a reduction of trainable parameters as well as result in increased training and test rates.

Tobias Schlosser, Frederik Beuth, Michael Friedrich et al.

Automated visual inspection in the semiconductor industry aims to detect and classify manufacturing defects utilizing modern image processing techniques. While an earliest possible detection of defect patterns allows quality control and automation of manufacturing chains, manufacturers benefit from an increased yield and reduced manufacturing costs. Since classical image processing systems are limited in their ability to detect novel defect patterns, and machine learning approaches often involve a tremendous amount of computational effort, this contribution introduces a novel deep neural network based hybrid approach. Unlike classical deep neural networks, a multi-stage system allows the detection and classification of the finest structures in pixel size within high-resolution imagery. Consisting of stacked hybrid convolutional neural networks (SH-CNN) and inspired by current approaches of visual attention, the realized system draws the focus over the level of detail from its structures to more task-relevant areas of interest. The results of our test environment show that the SH-CNN outperforms current approaches of learning-based automated visual inspection, whereas a distinction depending on the level of detail enables the elimination of defect patterns in earlier stages of the manufacturing process.