Find Academic Supervisors

Subhransu MajiView profile

Associate Professor

• Computer Vision
• Machine Learning
• Artificial Intelligence
• +5 more

Subhransu Maji is an Associate Professor in the Manning College of Information and Computer Sciences at the University of Massachusetts Amherst, and the co-director of the Computer Vision Lab. He is also affiliated with the Center for Data Science and holds a part-time role as an Amazon Scholar. His research focuses on high-level visual recognition algorithms and interdisciplinary applications in ecology and astronomy. He has received prestigious awards including the NSF CAREER Award (2018), Best Paper at WACV 2015, and the Google Graduate Fellowship (2008). Education: PhD in Computer Science from UC Berkeley (2011), BTech from IIT Kanpur (2006). Prior roles include Research Assistant Professor at Toyota Technological Institute at Chicago (2012-2014). Research Interests: Computer Vision Machine Learning AI Applications in Ecology and Astronomy 3D Shape Understanding Climate Science Grants and Funding: Supported by NSF, NASA, Climate Change AI, and industry grants from Facebook, NVIDIA, Adobe, and Dolby. Current projects include satellite imagery analysis for ecology and material science applications using deep learning. Labs and Teams: Leads the Computer Vision Lab, collaborates with interdisciplinary teams on ecological monitoring (e.g., bird migration tracking via radar data) and material property prediction (e.g., zeolite adsorption modeling).

• Machine Learning
• Deep Learning
• Theoretical Foundations of AI
• +11 more

Mikhail (Misha) Belkin is a Professor at the Halicioglu Data Science Institute (HDSI) at the University of California San Diego , with an affiliated appointment in the Department of Computer Science and Engineering . He is also an Amazon Scholar , reflecting his impactful industry collaboration. Since January 2024, he has served as the Editor-in-Chief of the SIAM Journal on Mathematics of Data Science (SIMODS) . Research Interests: Belkin's research centers on the theoretical foundations of machine learning, particularly the mathematical understanding of modern deep learning. His work investigates interpolation , over-parameterization , and feature learning in neural networks. He is renowned for introducing the double descent risk curve, which reconciles classical bias-variance trade-offs with the success of overfitted models. His recent work identifies the Average Gradient Outer Product (AGOP) as a fundamental mechanism of feature learning, applicable across architectures like CNNs and transformers. Scientific Contributions and Trends: His recent publications, appearing in Science , PNAS , and NeurIPS , demonstrate a strong trend toward unifying theories of generalization and optimization in over-parameterized systems. He explores how interpolating models can be statistically optimal, how gradient descent converges in non-convex landscapes via the PL* condition, and how kernel methods can be enhanced to perform feature learning. ACM Fellow (2023) Editor-in-Chief, SIAM Journal on Mathematics of Data Science (2024–present) Advising and Grants: Belkin actively mentors students and collaborators such as Adityanarayanan Radhakrishnan , Daniel Beaglehole , and Chaoyue Liu , who are frequent co-authors. He is a Principal Investigator (PI) in the Collaboration on the Theoretical Foundations of Deep Learning , funded by the NSF and Simons Foundation. He is also an external collaborator with the Eric and Wendy Schmidt Center at the Broad Institute and part of the NSF-funded TILOS AI Institute . Laboratories and Teams: While not explicitly named, his research group at UCSD is deeply involved in theoretical machine learning, focusing on the intersection of statistics, optimization, and deep learning. His work often involves large-scale collaborations and is closely tied to initiatives like SIMODS and TILOS.

• Digital Archaeology
• Computational Archaeology
• Machine Learning in Archaeology
• +13 more

Karsten Lambers is Professor of Digital and Computational Archaeology at the Faculty of Archaeology, Leiden University, where he leads research and teaching in the application of computational methods to archaeological data. His work integrates machine learning, remote sensing, text mining, and citizen science to advance archaeological prospection and heritage management. He is affiliated with the Department of Archaeological Sciences and plays key roles in research groups and university-wide initiatives such as SAILS and ARCHON. His research interests span Digital Archaeology , Machine Learning in Archaeology , Remote Sensing , Geoarchaeology , and Human-Environment Interaction . He investigates how computational tools can extract meaningful archaeological information from large datasets, including LiDAR imagery and excavation reports. His fieldwork spans Central Europe and Latin America, with a focus on prehistoric landscapes and cultural heritage. The analysis of his recent publications reveals a strong trend toward automated detection using deep learning (e.g., R-CNN, WODAN), named entity recognition in archaeological texts (e.g., ArcheoBERTje), and citizen science integration for data validation. His work bridges archaeology with computer science, geomatics, and environmental science, emphasizing interdisciplinary collaboration and methodological rigor. His scientific awards include: Best Thesis Award (University of Zurich, 2005) EUROPA NOSTRA Award (2020, 2022) Membership in the German Archaeological Institute (since 2022) Lambers actively supervises students and leads major research projects such as ABMA, EXALT, and Heritage Quest. He has secured substantial research funding and collaborates widely with computer scientists, geophysicists, and palaeoecologists. His teaching includes digital methods, modeling, and simulation, often linked to ongoing research. He has also contributed to open educational resources and digital textbooks in archaeology. He leads or participates in several research labs and teams, including the Digital Archaeology Research Group (which he chairs), the Heritage Quest citizen science project, and interdisciplinary teams focusing on alpine terraces and Iraqi prospection. His work emphasizes the integration of digital tools into practical archaeological workflows, advocating for complementary human-computer strategies.

Zhonghai LuView profile

Professor

• Computer Architecture
• Embedded Systems
• Network-on-Chip (NoC)
• +7 more

Zhonghai Lu is a Professor of Electronic Systems Design (specializing in Dependable and Autonomous Systems) at KTH Royal Institute of Technology, part of the Department of Electrical Engineering in the School of Electrical Engineering and Computer Science (EECS). He serves as Program Director for KTH's Embedded Systems master's program and Director of Studies at the Division of Electronics and Embedded Systems. His research focuses on Network-on-Chip (NoC), computer architecture, embedded systems, and Prognostics and Health Management (PHM) of power electronics. He leads a research group exploring in-network processing and embedded intelligence, transforming passive networks into active computational frameworks. Lu holds a BSc from Beijing Normal University (1989), MSc and PhD from KTH (2002, 2007), and an MBA in Innovation and Growth from the University of Turku (2012). He has authored over 240 scientific papers, including journal articles and peer-reviewed conferences, with notable recognitions such as Best Paper Awards at NOCS’2015 and EU HiPEAC, and a Featured Paper in IEEE Transactions on Computers (2020). He serves as Associate Editor for ACM Transactions on Architecture and Code Optimization (TACO) and has chaired major conferences like HiPEAC’2017 and NOCS’2018. His research group’s recent work includes integrating AI into hardware acceleration, fault-tolerant neural networks, and RUL estimation for power electronics using recurrent neural networks. Lu has secured grants from the Swedish Research Council and Intel Corporation and developed courses like IL2230 (Hardware Architectures for Deep Learning) and IL2233 (Embedded Intelligence), pioneering embedded AI education at KTH. Education: BSc (Beijing Normal University), MSc/PhD (KTH), MBA (University of Turku) Awards: Best Paper Awards (NOCS, EU HiPEAC), Swedish Research Council Grants, Intel Research Gifts Labs/Teams: Research Group on In-Network Processing and Embedded Intelligence

• Computer Architecture
• Machine Learning
• FPGA Acceleration
• +4 more

Hadi Esmaeilzadeh is an Associate Professor at the University of California, San Diego in the Department of Computer Science and Engineering . He previously served as an Assistant Professor at Georgia Tech (2013-2017) and holds the Halicioğlu Chair in Computer Architecture . As founder/director of the Alternative Computing Technologies (ACT) Laboratory and associate director of UCSD's Center for Machine Integrated Computing and Security (MICS) , his research drives cross-stack solutions for next-generation computer systems. Early tenure recipient at UCSD Coined the term "dark silicon" in computer architecture Developed Tabla/DnnWeaver open-source frameworks Research Interests span: Approximate Computing Neural Acceleration FPGA/ASIC Hardware Design Machine Learning Systems Dark Silicon Challenges Security/Privacy in Accelerated Systems Scientific Recognition : 4 CACM Research Highlights 4 IEEE Micro Top Picks Distinguished Paper Award (HPCA 2016) Inducted to ISCA Hall of Fame (2018) Teaching : Developed courses on accelerator design (CSE 240D) and alternative computing (CS 8803 ACT) Advocates for hands-on FPGA-based learning in Processor Design with FPGAs courses

Zhijian LiuView profile

Researcher

• Efficient Machine Learning
• Systems
• Sparse Convolution
• +2 more

Zhijian Liu is a Research Scientist at NVIDIA with a PhD from MIT, advised by Song Han. His work focuses on efficient machine learning and systems through sparse computation and hardware-aware neural network design. Rising Star in Data Science (UChicago/UCSD) Rising Star in ML and Systems (MLCommons) Qualcomm Innovation Fellowship awardee Research highlights include: SPVCNN++ for LiDAR segmentation HAQ framework in Intel OpenVINO TorchSparse framework for 3D CNN efficiency His recent publications explore: Efficient visual language models (NVILA) Contextual sparsity in LLM fine-tuning (SparseLoRA) Training-free acceleration of diffusion LLMs Query-aware visual sparsity mechanisms Contact: zhijian@mit.edu

Benyuan LiuView profile

Professor

• Data and Computer Communication Networks
• Mobile and Wireless Networks
• Internet Technologies & Applications
• +3 more

Benyuan Liu is a Professor at the Miner School of Computer and Information Sciences within the Kennedy College of Sciences at the University of Massachusetts Lowell . He serves as Director and Graduate Coordinator for Ph.D. programs, with expertise in Data and Computer Communication Networks, Mobile and Wireless Networks, and Internet Technologies & Applications. Education: B.S., University of Science and Technology of China M.S., Yale University Ph.D., University of Massachusetts Amherst His research focuses on Artificial Intelligence in Medical Imaging , Deep Learning for Endoscopy , and Edge Computing Systems . Recent work includes automated lesion detection, 3D reconstruction from sensor data, and predictive models for financial and reproductive health domains. The 15 most recent publications highlight applications of deep learning in medical diagnostics (thyroid nodules, gastric lesions, dental caries), computer vision (attention mechanisms, transformers), and financial technology (market psychology analysis). Technical themes include mmwave radar processing, diffusion models for synthetic data, and multi-scale feature extraction. Benyuan Liu leads the Computer Networking Lab and CHORDS initiative at UMass Center for Digital Health. His work bridges network optimization with healthcare AI , emphasizing real-time systems and portable diagnostics.

• Brain-Computer Interfaces
• Neurorehabilitation
• Haptic-User Interaction
• +4 more

Chang S. (CS) Nam is a Professor of Industrial and Systems Engineering at North Carolina State University, with affiliations in the UNC/NCSU Joint Department of Biomedical Engineering and the Department of Psychology. He holds editorial roles as Editor-in-Chief of Brain-Computer Interfaces journal and is a Fellow of the Human Factors and Ergonomics Society. His research focuses on advancing human-centered technologies through neuroergonomics, brain-computer interfaces (BCIs), and rehabilitation engineering, supported by grants from NSF, AFRL, AFOSR, and the Laboratory for Analytic Sciences. Education: Ph.D. in Industrial Engineering, Virginia Tech (2003) MSIE in Industrial Engineering, State University of New York at Buffalo (2000) Research Interests: His work spans neurorehabilitation, haptic interaction design, and BCI applications, aiming to bridge human cognitive and neural processes with technological systems. Key areas include adaptive human-computer interaction, EEG-based emotion recognition, and spatial augmented reality for skill training. Grants & Awards: Recipient of NSF CAREER Award (2010), Air Force Summer Faculty Fellowship (2018), and the Leland S. Kollmorgen Spirit of Innovation Award (2019) Editor of Brain-Computer Interfaces Handbook: Technological and Theoretical Advances (CRC Press) Advising & Labs: Advises students like Lingchao Mao and Priscille Koutouan Leads the Brain-Computer Interface and Neuroergonomics Lab, developing BCI systems for communication in vegetative states and neurorehabilitation

• Audio Signal Processing
• Spatial Audio Technology
• Medical Engineering
• +3 more

Craig Jin is an Associate Professor at the University of Sydney, leading the CARlab (Computing and Audio Research Laboratory) and Spatial Audio Research initiatives within the School of Electrical and Computer Engineering. He holds a BS from Stanford University, an MS from Caltech, and a PhD from the University of Sydney. His work focuses on immersive audio technologies, biomedical signal processing, and assistive technologies for sensory augmentation. Research interests include spatial audio reproduction, binaural processing, acoustic sensing for accessibility, and machine learning applications in signal processing. Key contributions span HRTF interpolation, noise reduction algorithms, and acoustic touch systems for the visually impaired. Recent projects include real-time MRI analysis of vocal tract dynamics and sparse recovery techniques for sound field reconstruction. His publications span over 150 peer-reviewed articles in journals like IEEE Transactions on Audio, Speech, and Language Processing, and conferences such as ICASSP. He advises four current PhD/Master’s students on projects like predictive gesture tracking, voice disorder classification, and magnetic resonance imaging techniques.

Rozenn DahyotView profile

Professor

• Computer Vision
• Machine Learning
• Digital Signal Processing
• +4 more

Rozenn Dahyot is a Professor of Computer Science at Maynooth University within the Faculty of Science & Engineering. She previously held roles as Assistant and Associate Professor in Statistics at Trinity College Dublin (2008-2021) and Lecturer in Computer Science (2005-2008). Her research interests bridge Digital Signal Processing, Computer Vision, Machine Learning, and Statistical Analysis. She organized the European Signal Processing Conference (EUSIPCO2021) in Dublin and served as President of the Irish Pattern Recognition and Classification Society (IPRCS) from 2014-2020. Her work spans topics like semantic scene understanding, CNN compression, and medical image segmentation. Key contributions include advancements in graph-based image analysis, reinforcement learning optimization, and AI-driven systems for disaster management. Dahyot is a member of IEEE, ACM, and EURASIP, contributing to both academic and industrial collaborations.

• Computational Pathology
• Medical Image Analysis
• Deep Learning
• +4 more

Nasir M. Rajpoot is a Professor in the Department of Computer Science at the University of Warwick, UK. His research focuses on computational pathology, medical image analysis, and deep learning applications in histology. He leads interdisciplinary projects integrating artificial intelligence with healthcare, particularly in cancer diagnostics and pathology workflows. Rajpoot’s work emphasizes developing robust algorithms for histology image analysis, including nuclear segmentation, tumor classification, and domain generalization in computational pathology. His contributions include the TIAToolbox, an open-source framework for tissue image analytics, and the CoNIC Challenge to advance nuclear detection and counting in histology images. He collaborates with clinicians and biologists to translate AI models into clinical practice, addressing challenges like tumor heterogeneity and staining variability. Rajpoot’s research spans colorectal, lung, and oral cancers, with a focus on predicting clinical outcomes via histological features and genomic data integration. Notable projects include the development of Handcrafted Histological Transformer (H2T) for unsupervised representations of whole slide images and the SAFRON framework for histology image synthesis. His work addresses domain adaptation, robustness evaluation, and explainability in AI-driven pathology systems.

Olga VekslerView profile

Professor

• Computer Vision
• Machine Learning
• Deep Learning
• +6 more

Olga Veksler is a Professor at the University of Waterloo's Department of Computer Science, part of the Faculty of Mathematics. She holds a Ph.D. and M.Sc. from Cornell University (1999) and a B.A. from New York University (1995). Her research focuses on computer vision, machine learning, and discrete optimization, with notable contributions to image segmentation, graph algorithms, and deep learning integration. Her work emphasizes semantic segmentation, salient object detection, and efficient optimization techniques for graphical models. Education: Ph.D. in Computer Science, Cornell University, 1999 M.Sc. in Computer Science, Cornell University, 1999 B.A. in Computer Science, New York University, 1995 Her research explores intersections between machine learning and traditional computer vision challenges, particularly leveraging graph-based optimization and CRF models. Recent trends in her work include weakly supervised learning, sparse non-local CRF applications, and test-time adaptation strategies for salient object detection. She has pioneered methods for shape priors in multi-object segmentation and efficient graph-cut algorithms. Her advising and grant activities are foundational to her research, though specific grant details are not listed here. She maintains a lab focused on advancing computer vision through algorithmic innovation, with contributions to both theoretical frameworks and practical applications in medical imaging and scene understanding.

Jaime S. CardosoView profile

Associate Professor

• Computer Vision
• Machine Learning
• Decision Support Systems
• +3 more

Jaime S. Cardoso is an Associate Professor with Habilitation at the Faculty of Engineering of the University of Porto (FEUP) and a Senior Researcher in the 'Information Processing and Pattern Recognition' Area at INESC TEC's Telecommunications and Multimedia Unit. He has been serving as Research Coordinator since September 15, 1998, and is a Senior Member of IEEE as well as co-founder of ClusterMedia Labs. His educational background includes a Licenciatura in Electrical and Computer Engineering (1999), an MSc in Mathematical Engineering (2005), and a Ph.D. in Computer Vision (2006), all from the University of Porto. Cardoso's research focuses on three major areas: computer vision, machine learning, and decision support systems. His work spans medical image analysis, explainable AI, semantic audio-visual analysis, and pattern recognition. He has co-authored over 150 papers, with more than 50 published in international journals, and has accumulated over 6,500 citations. His recent publications demonstrate expertise in cell nuclei segmentation, ordinal regression for CNNs, semantic segmentation with ordinal relationships, face recognition using synthetic data, and explainable vision language models for medical applications. Honorable Mention in the Exame Informática Award 2011 for 'Semantic PACS' First Place in the ICDAR 2013 Music Scores Competition Cardoso has supervised numerous graduate students at UP-FEUP, with recent theses focusing on multimodal explanations, autonomous driving, medical diagnosis, and explainable AI. His research group works at the intersection of computer vision, machine learning, and practical applications in healthcare and autonomous systems.

• Computer Engineering
• Embedded Systems
• Fault Tolerance
• +4 more

Corrado De Sio is a Fixed-term Researcher at the Department of Control and Computer Science (DAUIN) , Politecnico di Torino , affiliated with the College of Computer, Film, and Mechatronics Engineering . His academic roles include course instruction and collaboration for Reconfigurable Computing , High Performance Computing (HPC) , and Operating Systems for High-Performance Supercomputers across multiple academic years (2020-2025). Research Interests focus on: Reliability of reconfigurable systems and FPGAs under radiation effects Hardware-software co-design for fault tolerance Embedded systems in aerospace and safety-critical applications Machine learning acceleration on reconfigurable hardware Radiation effects on real-time operating systems and CNN implementations Recent publications address: 2025: Selective hardening of RISCV soft-processors for space applications 2025: Real-time 'signal for help' gesture recognition systems 2024: Reliability analysis of RISC-V processors and CNN placement algorithms 2023: Fault tolerance in FPGA-based CNNs and radiation effects on RTOS Patents include: PyXEL - Python Toolkit for Reconfigurable Hardware Surveillance Software for Real-Time Violence Detection Academic Supervision : Co-supervisor for PhD candidate Arash Amini Bardpareh in Computer and Systems Engineering .

Anne H Schistad SolbergView profile

Professor

• Machine Learning
• Deep Learning
• Image Analysis
• +6 more

Anne H Schistad Solberg is a Professor in the Department of Informatics at the University of Oslo's Faculty of Mathematics and Natural Sciences. She leads research in digital signal processing and image analysis, with a focus on machine learning applications across multiple domains. As co-director of SFI Visual Intelligence, she oversees research on interpretable deep learning models, uncertainty quantification, contextual learning, and self-supervised learning approaches. Her research spans medical imaging (particularly cardiovascular ultrasound), environmental monitoring using satellite imagery, and seabed mapping with sonar technology. Professor Solberg's work demonstrates a consistent trajectory from foundational signal processing techniques to cutting-edge deep learning applications. Her recent publications show increasing specialization in medical image analysis, particularly in echocardiography enhancement and cardiac structure segmentation, while maintaining strong contributions to remote sensing and geophysical applications. She teaches several popular courses including IN2070, IN3310, and IN5400 (Machine Learning for Image Analysis), which is noted as the most popular master's/PhD course on deep learning at the University of Oslo. Professor Solberg serves as principal investigator for the Intelligent Cardiovascular Ultrasound Scanner (INCUS) project, collaborating with GE Vingmed Ultrasound to develop AI-enhanced cardiac imaging systems that improve diagnostic accuracy and productivity in echocardiography. Co-director of SFI Visual Intelligence research center Principal Investigator for the INCUS project (Intelligent Cardiovascular Ultrasound Scanner) Member of the Digital Signal Processing and Image Analysis (DSB) research group Member of the Strategic Research Initiative: Multimodal Medical Imaging and Image Analysis (MEDIMA) Her research group develops algorithms that address real-world challenges in medical diagnostics and environmental monitoring, with a particular emphasis on making deep learning models more interpretable and reliable for critical applications. The INCUS project, funded through User-driven Research-based Innovation (BIA), aims to reduce the time wasted during cardiac ultrasound examinations by implementing intelligent algorithms that learn from expert users and historical data.