Doctoral School

Reading date: 09/12/2025

• SABRI ABREBEKOH, MOHAMMAD: Improving Efficiency of ReRAM-Based Accelerators for Cognitive Computing Workloads
  
  Author: SABRI ABREBEKOH, MOHAMMAD
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN COMPUTER ARCHITECTURE
  Department: Department of Computer Architecture (DAC)
  Mode: Normal
  Deposit date: 07/11/2025
  Reading date: 09/12/2025
  Reading time: 16:00
  Reading place: Sala d'Actes Edif. B6 - Planta baixa
  Thesis director: GONZÁLEZ COLÁS, ANTONIO MARIA | RIERA VILLANUEVA, MARC
  Thesis abstract: Deep Neural Networks (DNNs) have achieved remarkable success across a wide range of applications. The main operation in DNNs is the dot product between quantized input activations and weights. Previous works have proposed memory-centric architectures based on the Processing-in-Memory (PuM) paradigm. ReRAM technology is especially appealing for PuM-based DNN accelerators because of its high density for weight storage, low leakage energy, low read latency, and high-performance capabilities to perform DNN dot products massively in parallel within ReRAM crossbars. However, there are three main bottlenecks in ReRAM-based accelerators.First, the energy-hungry Analog-to-Digital Converter (ADC) required for in-ReRAM analog computations, which undermines the efficiency and performance benefits of PuM. To improve energy efficiency, we present ReDy, a hardware accelerator that implements a novel ReRAM-centric dynamic quantization scheme, leveraging bit-serial streaming and processing of activations. The energy consumption of ReRAM-based DNN accelerators is directly proportional to the numerical precision of input activations in each layer. ReDy exploits the fact that activations in convolutional layers are often grouped according to filter sizes and crossbar dimensions. It quantizes each group of activations on-the-fly with different precision levels, based on a heuristic that considers the statistical distribution of each group. Overall, ReDy significantly reduces ReRAM crossbar activity and the number of A/D conversions compared to static 8-bit uniform quantization. Evaluated on a set of modern CNNs, ReDy achieves on average 13% energy savings over an ISAAC-like accelerator, with negligible area overhead.Second, the costly writing process of ReRAM cells has led to accelerators designed with enough crossbar capacity to store entire DNN models. Given the continuous growth of DNN model sizes, this approach is infeasible for some networks and inefficient due to huge hardware requirements. These accelerators lack flexibility and face an adaptability challenge. To address this, we introduce ARAS, a cost-effective ReRAM-based accelerator that uses a smart scheduler to adapt various DNNs to resource-limited hardware. ARAS also overlaps computation of one layer with weight writing of others to mitigate high ReRAM write latency. Furthermore, ARAS introduces optimizations to reduce the energy overhead of ReRAM writes, including re-encoding weights to increase similarity across layers and reduce energy when overwriting cells. Overall, ARAS significantly reduces ReRAM write activity. Evaluated on multiple DNN models, ARAS delivers up to 2.2× speedup and 45% energy savings compared to a baseline PuM accelerator without optimizations, and up to 1.5× speedup and 62% energy savings compared to a TPU-like accelerator.Third, ReRAM cells suffer from limited endurance due to wear-out caused by repeated updates during inference, reducing the lifespan of ReRAM-based accelerators. Overcoming this endurance limitation is essential for making such accelerators viable in long-term, high-performance DNN inference. To address this, we propose Hamun, an approximate computing method designed to extend the lifespan of ReRAM-based accelerators through multiple optimizations. Hamun introduces a mechanism to detect and retire faulty cells caused by wear-out, preventing them from degrading accuracy. It also applies wear-leveling techniques across different abstraction levels and introduces a batch execution scheme to maximize cell utilization across inferences. Additionally, Hamun leverages the fault-tolerance of DNNs with a new approximation method that delays cell retirement, reducing the performance penalty and further extending lifespan. Evaluated on a set of DNNs, Hamun improves lifespan by 13.2× over a state-of-the-art baseline, with its main contributions coming from fault handling (4.6×) and batch execution (2.6×).
  

• SHEIKHSAMAD, MOHAMMAD: Learning Methods in Planning and Control for Autonomous Vehicles and Robotic Manipulation
  
  Author: SHEIKHSAMAD, MOHAMMAD
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN AUTOMATIC CONTROL, ROBOTICS AND VISION
  Department: Institute of Industrial and Control Engineering (IOC)
  Mode: Normal
  Deposit date: 11/11/2025
  Reading date: 09/12/2025
  Reading time: 12:00
  Reading place: Aula 28.8, Escola Tècnica Superior d'Enginyeria Industrial de Barcelona (ETSEIB), Av. Diagonal, 647, Planta 0, Pavelló G, 08028 Barcelona
  Thesis director: SUAREZ FEIJOO, RAUL | ROSELL GRATACOS, JOAN
  Thesis abstract: This thesis deals with the application of machine learning (ML) and deep learning (DL) techniques to enhance planning and control tasks across different fields, particularly on autonomous vehicles and robotic hands. Specifically, it addresses the learning-based development of a robust trajectory-tracking controller for autonomous vehicles, an adaptive path planner for robotic hands enabling dexterous manipulation, and a human-in-the-loop controller for myoelectric robotic hands to ensure precise grasp force regulation.In the field of autonomous vehicles, the thesis develops a Takagi–Sugeno (TS) controller using the Adaptive Neuro-Fuzzy Inference System (ANFIS) as a learning-based approach to infer a control strategy from input–output data of an existing controller. The closed-loop stability of the system is analyzed using Lyapunov theory and Linear Matrix Inequalities (LMIs). The proposed controller eliminates the need for online optimization, significantly reduces computational cost, and enhances real-time performance. Its effectiveness is validated through simulations on a small-scale autonomous vehicle.In the field of robotic dexterous manipulation, the thesis introduces three learning-based path planners using ANFIS and Deep Neural Networks (DNNs) to learn heuristics from an analytical planner and self-tune their parameters based on prior experience. This approach enables robots to manipulate objects of varying shapes, sizes, and material properties. The proposed planners are validated through real-world experiments using an Allegro robotic hand, demonstrating robustness against sensor noise and environmental disturbances.In the field of robotic grasping, the thesis presents a myocontrolled human-in-the-loop (HITL) system for precise grasp strength regulation. The system integrates both DNN-based and fuzzy-based force controllers. The fuzzy-based controller leverages fuzzy logic, with parameter optimization guided by user preferences collected through a graphical user interface (GUI) using Global Learning of Input–Output Strategies from Pairwise Preferences (GLISp). These controllers are compared against heuristic model-based controllers, and the system is validated through real-world experiments using the AR10 robotic hand, showing enhanced adaptability and fine-grained force regulation capabilities.The findings of this research contribute to the advancement of intelligent planning and control systems across multiple application areas, paving the way for more efficient, adaptive, and stable automation in real-world scenarios.
  

Reading date: 10/12/2025

• KHABBAZAN, BAHAREH: Improving Memory-centric Architectures for Accelerating Cognitive Computing Workloads
  
  Author: KHABBAZAN, BAHAREH
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN COMPUTER ARCHITECTURE
  Department: Department of Computer Architecture (DAC)
  Mode: Normal
  Deposit date: 31/10/2025
  Reading date: 10/12/2025
  Reading time: 15:00
  Reading place: Sala d'actes - Edif. B6 – Planta 0
  Thesis director: RIERA VILLANUEVA, MARC | GONZÁLEZ COLÁS, ANTONIO MARIA
  Thesis abstract: The rapid advancements in deep neural networks (DNNs) have led to increasingly complex and memory-intensive workloads, posing significant challenges for traditional computing architectures. Excessive data movement, computational inefficiencies, and energy constraints limit the scalability of DNN accelerators. This thesis addresses these challenges by proposing memory-centric approaches to optimize DNN execution through efficient quantization, in-memory processing, and data movement reduction.We first introduce DNA-TEQ, an adaptive exponential quantization scheme that minimizes memory footprint and eliminates the need for conventional multipliers, significantly enhancing energy efficiency. Experimental results show that DNA-TEQ reduces the memory footprint by 40% on average compared to the 8-bit integer baseline. The hardware processing-near-memory (PnM) accelerator designed to benefit from DNA-TEQ further improves inference latency by 1.5× while maintaining accuracy comparable to full-precision models.Next, we present QeiHaN, a PnM accelerator that employs base-2 exponential quantization and an implicit bit-shifting technique to reduce redundant memory accesses and optimize DNN inference. Our evaluations demonstrate that QeiHaN reduces memory movement by 67%, leading to a 4.2× speedup in execution time and a 3.5× reduction in energy consumption compared to conventional baseline architectures.Lastly, we propose Lama, a lightweight memory access mechanism that enhances lookup table (LUT)-based processing-in-memory (PuM) architectures by enabling parallel, column-independent accesses within DRAM mats, supporting up to 8-bit integer SIMD operations for large-scale models. The experimental results show that Lama significantly reduces memory commands for SIMD operations compared to the state-of-the-art PuM techniques. We further leverage Lama to design LamaAccel, an HBM-based large language model (LLM) accelerator that executes efficiently without modifying DRAM timing parameters. LamaAccel outperforms GPUs by up to 19×, achieving substantial energy savings in low-precision layers.The proposed techniques collectively reduce data movement, optimize memory utilization, and improve computational efficiency. Our findings demonstrate that memory-centric approaches can significantly enhance DNN acceleration, offering scalable and energy-efficient solutions for next-generation AI systems.
  

Reading date: 11/12/2025

• CRIOLLO ALIENDRES, CRUZ ARMANDO: Caracas Cinética: La transformación del paisaje urbano a partir de la inserción de obras de arte en los edificios públicos y privados, en los espacios públicos, la infraestructura vial y los sistemas masivos de transporte 1950-2012.
  
  Author: CRIOLLO ALIENDRES, CRUZ ARMANDO
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN URBANISM
  Department: Department of Urbanism, Territory and Landscape (DUTP)
  Mode: Normal
  Deposit date: 29/09/2025
  Reading date: 11/12/2025
  Reading time: 15:30
  Reading place: ETSAB (Esc. Técnica Sup. Arquit. Bcn)-Pl.Baja-Sala GradosAv. Diagonal, 649-651-08028-BCN(Videoconfencia: https://meet.google.com/ckz-quih-zjk-15:00)
  Thesis director: RUBERT DE VENTOS, MARIA
  Thesis abstract: This thesis analyzes the role of public art in the symbolic and social transformation of urban space in Caracas, with special emphasis on its transformative potential in environments marked by spatial fragmentation and a lack of public space. It is based on the premise that public art—particularly murals, sculptures, visual interventions, and ephemeral installations—intervenes in the relationships between citizens, territory, and collective memory.The research is based on a dual quantitative and qualitative approach, which articulates urban history, the cataloging and study of unique cases located in different urban environments (buildings, road infrastructure, the Metro, and the street), as well as an urban analysis from the 1950s to 2010. Emblematic cases are analyzed, such as works of art integrated into architecture, interventions linked to the network of avenues and highways, and monumental works such as those by Gego, Carlos Cruz Diez, Jesús Soto, and Alejandro Otero.The findings reveal that public art in Caracas serves multiple functions: it redefines urban spaces, reinforces local and metropolitan identities, and democratizes access to culture. The research identifies how the recurring practice of integrating art, architecture, and the city has evolved into an urban tradition that continues to this day, in an environment that poses tensions between visual art, urban policies, and the processes of appropriation of public space. Thus, art located in urban spaces plays a connecting role between institutional programs and social actors.Finally, the thesis compiles and organizes a section of the city's urban evolution, in which public art served as a catalyst for a more just, plural, and participatory city.
  

• DEL POZO MARTÍN, JORGE: Estudio estadístico del control de calidad del hormigón
  
  Author: DEL POZO MARTÍN, JORGE
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN CONSTRUCTION ENGINEERING
  Department: Department of Civil and Environmental Engineering (DECA)
  Mode: Normal
  Deposit date: 18/11/2025
  Reading date: 11/12/2025
  Reading time: 10:00
  Reading place: C1-002
  Thesis director: AGUADO DE CEA, ANTONIO | PIALARISSI CAVALARO, SERGIO HENRIQUE
  Thesis abstract: This doctoral thesis addresses a critical analysis of the structural concrete quality control system in Spain. Currently, national regulations establish a dual control system: one for production, executed by the manufacturer at the plant, and another for reception, carried out on-site upon receipt of the concrete. This duplication generates operational, technical, and economic conflicts, as well as potential inconsistencies in test results, which raises a debate about its suitability and effectiveness.The general objective of the first part of this thesis is to evaluate the efficiency and reliability of the dual control system, in order to subsequently propose an optimized model that simplifies the process without compromising structural safety or concrete quality. The aim is to move toward a more rational control adapted to the technological reality and European regulations.The second line of research focuses on the statistical basis on which the regulatory criteria for the acceptance or rejection of a batch of concrete are based. Currently, Spanish regulations assume that compression test results follow a normal (Gaussian) distribution. However, this hypothesis lacks solid theoretical justification and has limitations such as the possibility of obtaining negative values and a symmetry that does not always fit the real data. Therefore, other distribution functions are explored, such as the log-normal and Weibull distribution functions, which could better fit the actual results obtained in tests.Throughout the document, a methodology based on the analysis of large volumes of test data from real-life construction projects is presented. Different distribution functions are contrasted using goodness-of-fit tests, and the differences in the estimate of the 5% percentile, which defines the characteristic strength of concrete, are quantified. The results indicate that the normal function is not the most appropriate distribution function to best fit the data.Based on the findings obtained, the thesis proposes a review of the current quality control model, opting for a system based primarily on production control—with the possibility of receiving control using other types of tests that provide information about the finished structure—provided that traceability and quality are guaranteed through strict procedures and certifications. Likewise, it is suggested that normative statistical models be updated, incorporating distribution functions that more accurately represent the actual behavior of concrete.In conclusion, this research proposes a significant improvement in the way structural concrete quality is controlled in Spain. It provides technical, regulatory, and statistical foundations that justify a shift toward a more efficient model, free of redundancies, aligned with European guidelines, and supported by more robust statistical analysis, which could represent