Sepideh Pashami

ORCID: 0000-0003-3272-4145
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Fault Detection and Control Systems
  • Anomaly Detection Techniques and Applications
  • Advanced Chemical Sensor Technologies
  • Machine Learning and Data Classification
  • Machine Fault Diagnosis Techniques
  • Explainable Artificial Intelligence (XAI)
  • Traffic Prediction and Management Techniques
  • Insect Pheromone Research and Control
  • Machine Learning in Healthcare
  • Domain Adaptation and Few-Shot Learning
  • Time Series Analysis and Forecasting
  • Machine Learning and ELM
  • Neural Networks and Applications
  • Reliability and Maintenance Optimization
  • Data Mining Algorithms and Applications
  • Data Quality and Management
  • Imbalanced Data Classification Techniques
  • Adversarial Robustness in Machine Learning
  • Data Visualization and Analytics
  • Data Management and Algorithms
  • Traffic control and management
  • Statistical Methods and Inference
  • Infrastructure Maintenance and Monitoring
  • Quality and Safety in Healthcare
  • Social Robot Interaction and HRI

RISE Research Institutes of Sweden
2021-2025

Halmstad University
2016-2024

Örebro University
2011-2014

University of Tehran
2006

Smart cities and communities (SCC) constitute a new paradigm in urban development. SCC ideate data-centered society aimed at improving efficiency by automating optimizing activities utilities. Information communication technology along with Internet of Things enables data collection the help artificial intelligence (AI) situation awareness can be obtained to feed actors enriched knowledge. This paper describes AI perspectives gives an overview AI-based technologies used traffic enable road...

10.3390/smartcities4020040 article EN cc-by Smart Cities 2021-05-18

Evolutionary Algorithms (EAs) are often challenging to apply in real-world settings since evolutionary computations involve a large number of evaluations typically expensive fitness function. For example, an evaluation could training new machine learning model. An approximation (also known as meta-model or surrogate) the true function can be used such applications alleviate computation cost. In this paper, we propose two-stage surrogate-assisted approach address computational issues arising...

10.1016/j.eswa.2022.118528 article EN cc-by Expert Systems with Applications 2022-08-18

In classification problems, as the number of classes increases, correctly classifying a new instance into one them is assumed to be more challenging than making same decision in presence fewer classes. The essence problem that using learning algorithm on each boundary individually better several simultaneously. However, why and when it happens still not well-understood today. This work's main contribution introduce concept heterogeneity boundaries an explanation this phenomenon. Based...

10.1109/access.2022.3192514 article EN cc-by IEEE Access 2022-01-01

Domain adaptation (DA) methods facilitate cross-domain learning by minimizing the marginal or conditional distribution shift between domains. However, is not well addressed existing DA techniques for regression task. In this paper, we propose Multi-Domain Adaptation Regression under Conditional (DARC) method. DARC constructs a shared feature space such that linear on top of generalizes to all other words, aligns different domains according task-related information encoded in values dependent...

10.1016/j.eswa.2023.119907 article EN cc-by Expert Systems with Applications 2023-03-20

We address the problem of detecting changes in activity a distant gas source from response an array metal oxide (MOX) sensors deployed open sampling system. The main challenge is turbulent nature dispersion and dynamics sensors. propose change point detection approach evaluate it on individual experimental setup where intensity, compound, or mixture ratio. also introduce efficient sensor selection algorithm with selected subsets.

10.3390/s121216404 article EN cc-by Sensors 2012-11-27

Many industries today are struggling with early the identification of quality issues, given shortening product design cycles and desire to decrease production costs, coupled customer requirement for high uptime. The vehicle industry is no exception, as breakdowns often lead on-road stops delays in delivery missions. In this paper we consider issues be an unexpected increase failure rates a particular component; those particularly problematic original equipment manufacturers (OEMs) since they...

10.3390/info11070354 article EN cc-by Information 2020-07-06

We propose a two-stage surrogate-assisted evolutionary approach to address the computational issues arising from using Genetic Algorithm (GA) for feature selection in wrapper setting large datasets. The proposed involves constructing lightweight qualitative meta-model by sub-sampling data instances and then this carry out task.

10.1145/3583133.3595823 article EN 2023-07-15

This paper presents a comprehensive empirical investigation into the interactions between various randomization techniques in Deep Neural Networks (DNNs) and their impact on learning performance. It is well-established that injecting randomness training process of DNNs, through approaches, at different stages, often beneficial for reducing overfitting improving generalization. Nonetheless, such as weight noise, dropout, many others remain poorly understood. Consequently, it challenging to...

10.1016/j.ins.2024.120500 article EN cc-by Information Sciences 2024-03-21

10.1109/wacv61041.2025.00643 article EN 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025-02-26

Explainable Artificial Intelligence (XAI) fills the role of a critical interface fostering interactions between sophisticated intelligent systems and diverse individuals, including data scientists, domain experts, end-users, more. It aids in deciphering intricate internal mechanisms ``black box'' Machine Learning (ML), rendering reasons behind their decisions more understandable. However, current research XAI primarily focuses on two aspects; ways to facilitate user trust, or debug refine ML...

10.48550/arxiv.2306.05120 preprint EN cc-by arXiv (Cornell University) 2023-01-01

Mobility as a Service (MaaS) combines various modes of transportation to present mobility services travellers based on their transport needs. This paper proposes knowledge-based framework Artificial Intelligence (AI) integrate data types and provide with customized services. The proposed includes knowledge acquisition process extract structure from multiple sources information (such experts weather data). It also adds new base improves the quality previously acquired knowledge. We discuss...

10.3390/su15032717 article EN Sustainability 2023-02-02

Predictive Maintenance (PM) is a proactive maintenance strategy that tries to minimize system’s downtime by predicting failures before they happen. It uses data from sensors measure the component’s state of health and make forecasts about its future degradation. However, existing PM methods typically focus on individual measurements. While it natural assume history measurements carries more information than single one. This paper aims at incorporating such into models. In practice,...

10.3390/app10010069 article EN cc-by Applied Sciences 2019-12-20

Ensemble learning methods combine multiple models to improve performance by exploiting their diversity. The success of these approaches relies heavily on the dissimilarity base forming ensemble. This diversity can be achieved in many ways, with well-known examples including bagging and boosting. It is within an ensemble that allows correct errors made its members, consequently leads higher classification or regression performance. A mistake a model only rectified if other members behave...

10.1016/j.neunet.2021.03.002 article EN cc-by Neural Networks 2021-03-14

Using hierarchies of classes is one the standard methods to solve multi-class classification problems. In literature, selecting right hierarchy considered play a key role in improving performance. Although different have been proposed, there still lack understanding what makes good and method extract perform better or worse. To this effect, we analyze compare some most popular approaches extracting hierarchies. We identify common pitfalls that may lead practitioners make misleading...

10.1016/j.patcog.2022.109225 article EN cc-by Pattern Recognition 2022-11-30

Research on machine activity recognition (MAR) is drawing more attention because MAR can provide productivity monitoring for efficiency optimization, better maintenance scheduling, product design improvement, and potential material savings. A particular challenge of human-operated machines the overlap when transiting from one to another: during transitions, operators often perform two activities simultaneously, e.g., lifting fork already while approaching a rack, so exact time ends another...

10.1016/j.engappai.2023.106992 article EN cc-by Engineering Applications of Artificial Intelligence 2023-08-28

Many applications of metal oxide gas sensors can benefit from reliable algorithms to detect significant changes in the sensor response. Significant indicate a change emission modality distant source and occur due sudden concentration or exposure different compound. As consequence turbulent transport relatively slow response recovery times sensors, their open sampling configuration exhibits strong fluctuations that interfere with interest. In this paper we introduce TREFEX, novel point...

10.3390/s130607323 article EN cc-by Sensors 2013-06-04
Coming Soon ...