Neural networks have been successfully applied in various resource-constrained edge devices, where usually central processing units (CPUs) instead of graphics processing units exist due to limited power availability. State-of-the-art research still focuses on efficiently executing enormous numbers of multiply-accumulate (MAC) operations. However, CPUs themselves are not good at executing such mathematical operations on a large scale, since they are more suited to execute control flow logic, i.e., computer algorithms. To enhance the computation efficiency of neural networks on CPUs, in this paper, we propose to convert them into logic flows for execution. Specifically, neural networks are first converted into equivalent decision trees, from which decision paths with constant leaves are then selected and compressed into logic flows. Such logic flows consist of if and else structures and a reduced number of MAC operations. Experimental results demonstrate that the latency can be reduced by up to 14.9 % on a simulated RISC-V CPU without any accuracy degradation. - The code is open source at https://github.com/TUDa-HWAI/NN2Logic

Compound words are a grammatical structure that allows forming new words by composing existing words. For e-commerce search in German, it is essential to split these compounds into meaningful parts because item titles often use the joint form while search queries are often split. We propose a method for German compound splitting leveraging a large language model (LLM) with a voting mechanism and a hyperparameter search for automatically optimizing prompt and parameter combinations. Our evaluation of the proposed method on human-created gold standard data for e-commerce shows that it outperforms existing methods for compound splitting in this domain.

In the realm of e-commerce, the process of search stands as the primary point of interaction for users, wielding a profound influence on the platform's revenue generation. Notably, spelling correction assumes a pivotal role in shaping the user's search experience by rectifying erroneous query inputs, thus facilitating more accurate retrieval outcomes. Within the scope of this research paper, our aim is to enhance the existing state-of-the-art discriminative model performance with generative modelling strategies while concurrently addressing the engineering concerns associated with real-time online latency, inherent to models of this category. We endeavor to refine LSTM-based classification models for spelling correction through a generative fine-tuning approach hinged upon pre-trained language models. Our comprehensive offline assessments have yielded compelling results, showcasing that transformer-based architectures, such as BART (developed by Facebook) and T5 (a product of Google), have achieved a 4% enhancement in F1 score compared to baseline models for the English language sites. Furthermore, to mitigate the challenges posed by latency, we have incorporated model pruning techniques like no-teacher distillation. We have undertaken the deployment of our model (English only) as an A/B test candidate for real-time e-commerce traffic, encompassing customers from the US and the UK. The model attest to a 100% successful request service rate within real-time scenarios, with median, 90th percentile, and 99th percentile (p90/p99) latencies comfortably falling below production service level agreements. Notably, these achievements are further reinforced by positive customer engagement, transactional and search page metrics, including a significant reduction in instances of search results page with low or almost zero recall. Moreover, we have also extended our efforts into fine-tuning a multilingual model, which, notably, exhibits substantial accuracy enhancements, amounting to a minimum of 16%, across four distinct European languages and English.

Finding the optimal decoding parameters in speech recognition is often done manually in a rather tedious manner, although automatic gradient-free optimization techniques have been shown to perform quite well for this task. While there have been recent scientific contributions in this field, no thorough comparison of possible methods, in terms of convergence speed and performance, has been undertaken. In this paper, we conduct a series of experiments with three decoding paradigms and four different optimization techniques found in recent literature, both on unconstrained and time-constrained decoder optimization. We offer our findings on the German Difficult Speech Corpus and on the LinkedTV test sets.

This paper introduces a framework for establishing links between related media fragments within a collection of videos. A set of analysis techniques is applied for extracting information from different types of data. Visual-based shot and scene segmentation is performed for defining media fragments at different granularity levels, while visual cues are detected from keyframes of the video via concept detection and optical character recognition (OCR). Keyword extraction is applied on textual data such as the output of OCR, subtitles and metadata. This set of results is used for the automatic identification and linking of related media fragments. The proposed framework exhibited competitive performance in the Video Hyperlinking sub-task of MediaEval 2013, indicating that video scene segmentation can provide more meaningful segments, compared to other decomposition methods, for hyperlinking purposes.

S.622-626

This paper describes the content of the video for ICRA 2010. The approach presented in the video is a complete automatic registration of all needed devices for robotic guided CO2 bone processing. The system consists of an optical tracking system, a lightweight robot and a scan head. While a standard point to point transformation for the transformation between optical tracking system and robot is inadequate, a special registration algorithm was developed and therefore is used. This allows an automatic registration by moving the robot to different positions and storing the position of a body attached to the robot TCP. The scan head is registered to the optical tracking system or to the robot using a camera that tracks the prototype laser. Only the patient must be registered manually. The result is that the robot moves to the target position on the bone defined inside the segmented 3D CT dataset. With the known registration between robot and tracking system the robot can also be controlled via visual servoing.

This paper presents a successful technique for creating a new language model (LM) that adapts the original target LM used by a machine translation (MT) system.This technique is especially useful for situations where there are very scarce resources for training the target side (Spanish Sign Language (LSE) in our case) in order to properly estimate the target LM, the Sign Language Model (SLM), used by the MT system.The technique uses information from the source language, Spanish in our task, and from the phrase-based translation matrix in order to create a new LM, estimated using web frequencies, which adapts the counts of the SLM through the Maximum A Posteriori method (MAP).The corpus consists of common used sentences spoken by an officer when assisting people in applying for, or renewing, the National Identification Document.The proposed technique allows relative reductions of 15.5% on perplexity and 2.7% on WER for translation, which are close to half the maximum performance obtainable when only the LM is optimized.