The proposed Bayesian optimization-based approach enhances micromixer performance by optimizing geometric parameters, significantly reducing required number of simulations, and accelerating the design process compared to conventional methods.

The paper reports a solution for the integration of industrial robot ABB IRB140 with system automatic speech recognition (ASR) and computer vision. has task to manipulate objects placed randomly on pad lying table, vision recognize their characteristics (shape, dimension, color, position, orientation). ASR human use it as command robot, so can objects.

Robotic systems for research and development of factory automation are complex unavailable broad deployment in robotic laboratory settings. The usual setup consists series sensors, arms mobile robots integrated orchestrated by a central information system. Cloud-based integration has been gaining traction recent years. In order to build such system environment, there several practical challenges that have be resolved come point when can become operational. this paper, we present the one...

This paper reports a spoken natural language dialogue system that manages the interaction between user and industrial robot ABB IRB 140. To extent is multimodal, it uses three communication modalities: (i) (automatic speech recognition text-to-speech synthesis), (ii) visual of figures determination their positions, (iii) typed text. adaptive, takes verbal spatial contexts into account in order to adapt its behavior process spontaneously formulated commands different syntactic forms without...

This paper reports on a pilot corpus of child-robot interaction in therapeutic settings. The comprises recordings the interactions between twenty-one children and conversational humanoid robot MARKO, kinesitherapeutic room at Clinic Paediatric Rehabilitation Novi Sad, Serbia. subject group included both healthy with cerebral palsy similar movement disorders. Approximately 156 minutes session time was recorded. All dialogues were transcribed, nonverbal acts annotated. initial evaluation...

The paper presents a system for stress detection in broiler chickens using audio data. It should be part of the precision livestock farming whose goal is to maximize profit and product quality without harming animal wellbeing. consisted 4 classifiers adapted age groups (one each week). These are based on support vector machines as input features they use voice evaluation speech emotion recognition. Accuracy 50ms frame level these varies from 63 83 %, depending group.

Non-adversarial robustness, also known as natural is a property of deep learning models that enables them to maintain performance even when faced with distribution shifts caused by variations in data. However, achieving this challenging because it difficult predict advance the types may occur. To address challenge, researchers have proposed various approaches, some which anticipate potential shifts, while others utilize knowledge about already occurred enhance model generalizability. In...

Emphasizing a data-centric AI approach, this paper introduces novel two-stage active learning (AL) pipeline for automatic speech recognition (ASR), combining unsupervised and supervised AL methods. The first stage utilizes by using x-vectors clustering diverse sample selection from unlabeled data, thus establishing robust initial dataset the subsequent AL. second incorporates strategy, with batch method specifically developed ASR, aimed at selecting informative batches of samples. Here,...

Distributed learning and inference algorithms have become indispensable for IoT systems, offering benefits such as workload alleviation, data privacy preservation, reduced latency. This paper introduces an innovative approach that utilizes unmanned aerial vehicles (UAVs) a coverage extension relay environmental monitoring in rural areas. Our method integrates split (SL) strategy between edge devices, UAV server to enhance adaptability performance of mechanisms. By employing UAVs by...

This paper introduces a cognitively-inspired symbolic framework for knowledge representation in human-machine interaction. The is developed within the ongoing research on computational model of hierarchical associative long-term memory. integrates neurocognitive understanding human memory system with selected insights from linguistics, and primarily addresses storage aspect proposed structure conceptualized as set (multisource-multisink) semantic flow networks, including units different...

Nonlinear state estimation (SE), with the goal of estimating complex bus voltages based on all types measurements available in power system, is usually solved using iterative Gauss-Newton (GN) method. The nonlinear SE presents some difficulties when considering inputs from both phasor measurement units and supervisory control data acquisition system. These include numerical instabilities, convergence time depending starting point method, quadratic computational complexity a single iteration...

Fifth-Generation (5G) networks have a potential to accelerate power system transition flexible, softwarized, data-driven, and intelligent grid. With their evolving support for Machine Learning (ML)/Artificial Intelligence (AI) functions, 5G are expected enable novel data-centric Smart Grid (SG) services. In this paper, we explore how data-driven SG services could be integrated with ML/AI-enabled in symbiotic relationship. We focus on the State Estimation (SE) function as key element of...

When training language models (especially for highly inflective languages), some applications require word clustering in order to mitigate the problem of insufficient data or storage space.The goal is group words that can be well represented by a single class sense probabilities appearances different contexts.This paper presents comparative results obtained using approaches when N-gram Serbian, as based on recurrent neural networks.One approach unsupervised optimized Brown's algorithm, which...

Speech recognition systems are commonly modelled by hidden Markov models with Gaussian mixture as observation density functions. These have a significant number of parameters, which usually leads to the problem data sparsity, especially for under-resourced languages such Serbian. One ways overcome sparsity is reduction features. Linear discriminant analysis (LDA) and heteroscedastic LDA (HLDA) two common reduce dimensionality in an automatic speech task. The paper compares properties Serbian...