Thesis offers

A list of issues hard to compute would include:

1. The computation of the partition function is NP-hard, involving an  exponential sum of terms
2. The exact computation of the derivative of the log-likelihood is  also NP-hard, since it contains the derivative of the partition function

Therefore, in practice we have to approximate both the computation of
the probabilities and several components of the learning process
itself. These drawbacks have prevented RBMs to show their real
potential as truly probabilistic models.

Currently, we are working on trying to improve several of the unsolved
issues related to RBMs:

1. - Mechanisms to control the learning process: www.lsi.upc.edu/~eromero/Publications/Downloads/2018-tnnls-stopcritRBM.pdf
2. Better approximations of the derivative of the log-likelihood: http://www.lsi.upc.edu/~eromero/Publications/Downloads/2019-nn-weightedCD.pdf
3. Efficient approximation of the partition function (work in progress)

These works have opened new lines of research, some of which can be
the topic of a Master's Thesis. The scope and degree of depth of the
work can be adapted to the estimated times to complete the Thesis. For
further details, contact Enrique Romero (eromero@cs.upc.edu).

Quantum computers promise exponential improvements over conventional ones due to the extraordinary properties of qubits. However, a quantum computer faces many challenges relative to the movement of qubits which is completely different from the movement of classical data. This thesis delves into these challenges and proposes solutions to create scalable quantum computing systems and the algorithms that run within them, following the current European projects at N3Cat (www.n3cat.upc.edu) on scalable quantum computing.

The interested candidate will work in a group of several PhD students and in collaboration with Universitat Politècnica de València, working in ONE of the following areas:

• Study the use of AI to model quantum computers or to assist in the development of architectural methods (qubit mapping, qubit partitioning, qubit routing).
• Study the potential scaling and adaptation of the existing quantum computing architectures to specific quantum machine learning algorithms (QAOA, VQE, others).

 The goal of this thesis is to enable the application of Deep Reinforcement Learning techniques for optimization problems in very large and complex scenarios, in this case Data Center Networks. Inspired by the last related work from OpenAI (https://openai.com/blog/evolution-strategies/), we will explore time-efficient alternatives to train a Graph Neural Network based DRL agent on large optimization scenarios in a reasonable amount of time.

A more detailed description of the project can be found in

https://www.cs.upc.edu/~eromero/Downloads/Retina-TFM-Project-01.pdf

The project is proposed in collaboration with Javier Zarranz Ventura
(Institut Clínic d'Oftalmologia, ICOF, Hospital Clínic de Barcelona, and
Institut d'Investigacions Biomèdiques August Pi I Sunyer, IDIBAPS),
which would provide a large annotated database to develop the project. For further information, please contact Alfredo Vellido (avellido@cs.
upc.edu) or Enrique Romero (eromero@cs.upc.edu).

The critical privacy issues underscore the urgent need for more robust oversight, regulation, and structural incentives. These measures are crucial to safeguard citizens' medical data and to address the technical challenges of maintaining anonymity and preventing re-identification in the face of increasingly powerful AI-based algorithms.
This project is dedicated to ensuring the benefits of AI-based systems in healthcare are balanced with the rigorous data protection standards mandated. Our commitment to careful planning, robust technical controls, and a human-centric approach is a must, ensuring that our approach leverage AI-based systems while safeguarding individuals' privacy.
Some of the  most pressing and relevant privacy issues when using AI-based methods to handle medical data include:
Data Access, Use, and Control
Breach of Patient Privacy
Lack of Oversight and Regulation
Competing Interests and Incentives
Biased and Inaccurate AI-based Models

The project is a collaboration with cardiologists from Hospital Clínic de Barcelona that will provide the data and the specialized validation.

Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging technique that measures changes in oxygenated (HbO2) and deoxygenated hemoglobin (HbR) in the cerebral cortex. Due to its portability and low cost, fNIRS has been used in Brain-Computer Interface (BCI) applications, characterizing hemodynamic responses to varying stimuli, and investigating auditory and visual-spatial attention during Complex Scene Analysis (CSA). In this project, we want to design and implement an fNIRS study with a goal of studying the impact of neural and BCI outcomes to improving the training of LAI models' attention mechanism (e.g., Transformer attention) during reading comprehension tasks (e.g., the participants will be judging the quality of generated text). We want to demonstrate experimentally that augmenting a model with fNIRS data carries neural activity features complementing the information captured by the model and demonstrate that it improves the models' performance. To this end, we will have to collect data from participants and test how different Transformer models benefit from different types of fNIRS attention masks.

The candidate will:

• Carry out controlled studies and collect data from a number of participants.
• Try different approaches to incorporating fNIRS signal in the training process of DL/LLM models and compare against baselines/experiments (these should be replicated).
• Carry out ablation studies, personalisation techniques, error analyses, etc.
• Contribute to authoring a scientific article

Els detalls es donaran a l'alumne personalment

Immobilitzar objectes, ja sigui agafant-los amb una mà robòtica amb dits o utilitzant qualsevol dispositiu de fixació, és un camp de recerca força actiu en robòtica i automatització. L'objectiu és restringir el possible moviment de l'objecte malgrat la possible existència de pertorbacions externes que actuen sobre ell (per descomptat, fins a una magnitud determinada).

En termes generals, els objectes es poden classificar en dos tipus: rígids i deformables. El treball a realitzar en aquest projecte tracta el cas particular de objectes deformables composts per un conjunt d'enllaços rígids units entre si per qualsevol tipus d'articulació que ha de ser immobilitzat en la seva configuració actual.

Les Regions de Contacte Independents (ICRs) són regions de la superfície de l'objecte tal que un contacte amb un dit en qualsevol punt de cadascuna d'elles garanteix una prensió que  immobilitza l'objecte independentment de la posició exacta del contacte.

L'objectiu específic d'aquest projecte és implementar un paquet de programari que permeti el càlcul d'ICRs donat el model i la configuració de l'objecte articulat a immobilitzar.

Alguns supòsits i condicions específiques:

• Es presumeix que els objectes estan descrits per un núvol de punts, i es coneix la direcció normal a la superfície de l'objecte en cada punt. El núvol de punts pot generar-se a partir d'un model d'objecte (per exemple, CAD) o d'una càmera 3D.

• Es prefereix l'ús de C++ o Python.

Publicacions científiques relacionades:

•  N. Alvarado, R. Suárez and M. Roa, ''Determining independent contacts regions to immobilize 2D articulated objects'' 2015 IEEE International Conference on Robotics and Automation, ICRA 2015, Seattle, Washington, USA, May 26-30, 2015, pp. 4292-4297. (DOI 10.1109/ICRA.2015.7139791).

Aquest article presenta un treball equivalent per a objectes en 2D, el treball a realitzar és la implementació del mateix enfocament però en 3D.

• M. Roa and R. Suárez, ''Computation of independent contact regions for grasping 3D objects'', IEEE Transactions on Robotics (ISSN: 1552-3098), Volume 25, Issue 4, Aug. 2009, Pages 839-850.

Aquest article presenta la idea clau de l'enfocament per a objectes sòlids en 3D (no articulats), el treball a realitzar és la implementació d'una generalització de l'enfocament per a objectes articulats.

Posar-se en contacte amb raul.suarez@upc.edu per a més informació.

Recent advancements in nanotechnology have enabled the development of means for sensing and wireless communications with unprecedented miniaturization and capabilities, to the point that they can be introduced into the gastrointestinal tract inside a pill or into the bloodstream in the form of passively flowing nanomachines. 

This opens the door to the idea of intra-body communication networks, this is, a swarm of nanosensors inside the human body that use communications to coordinate their actions to sense and localize specific events (lack of oxygen, biomarkers, etc). This can lead to the development of applications such as continuous monitoring of diabetes, detection and localization of cancer micro-tumors, or early detection of blood clots. These possibilities are currently investigated by our team at the N3Cat (www.n3cat.upc.edu).

In this context, we are looking for excellent and self-motivated individuals who are eager to work on developing AI schemes (based on graph neural networks or multi-agent RL) for the detection and localization of events inside of the human body. Data will be gathered with an in-house simulator that integrates mobility models (BloodVoyagerS) and communication models (TeraSim).

The application of Artificial Intelligence (AI) and Machine Learning (ML) to network security (AI4SEC) is paramount against cybercrime. While AI/ML is mainstream in domains such as computer vision and natural language processing, traditional AI/ML has produced below-par results in AI4SEC. Solutions do not properly generalize, are ineffective in real deployments, and are vulnerable to adversarial attacks. A fundamental limitation is the lack of AI/ML technology specific to network security. Due to their unique ability to learn and generalize over graph-structured information, graph-learning approaches, and in particular Graph Neural Networks (GNNs), have recently enabled groundbreaking applications in multiple fields where data are generally represented as graphs. Network security data are intrinsically relational, and initial research suggests that graph-structured representations and GNNs have the potential to become foundational to AI4SEC, in the way convolutional and recursive networks were to computer vision and natural language processing.
The goal of GRAPHS4SEC is to leverage graph data representations and modern GNN technology to conceive a new breed of robust GNN-based network security methods which could radically advance the AI4SEC practice. The objectives of GRAPHS4SEC are: (a) to investigate algorithmic methods that facilitate modeling and learning from graph-based network security data; (b) to compare the benefits and overheads of GNN-based AI4SEC to traditional AI/ML in terms of detection performance, generalization, scalability, and robustness against adversarial attacks; (c) to showcase the benefits and improvements of GRAPHS4SEC technology in four critical, real-world network security applications with significant impact for society, considering (in particular) the detection and early mitigation of phishing and fake/malicious websites, a threat among the most popular and society-wide harmful in today's Internet.

In the proposed project, we are interested in the mobile setting, and propose the use of depth information, on top of the usual RGB (Red, Green, Blue) pixel data acquired by mobile device cameras, to track and quantify visual attention. Depth information can either be estimated with post computer visual processing or acquired directly with high precision, using advanced LiDAR sensors that newer smartphone models are currently equipped with. Furthermore, we propose the use of "lightweight" Deep Learning (DL) techniques to achieve an accurate and robust gaze point estimation without the need of extensive calibration. Unlike prior art, we will consider the use of historical information to inform future predictions and will examine how such models can be tailored to each user by injecting personalized data. In addition, we propose a more lightweight Neural Network (NN) architecture, suitable for the processing power of mobile devices, that also leverages the use of video or photo frames with depth sensory data.

The candidate will:

• Develop an app for iPhone (in Swift) for collecting gaze information and image data (using the device's frontal camera and LiDAR sensors), while completing a simple task of tracking a red dot in a matrix.
  
• Incorporate in the app the functionality offered by SensingKit,  an open-source mobile sensing framework compatible with iOS specialised in capturing motion, orientation, location and proximity (using iBeacon¿ or Eddystone¿) data and transmitting them on a server over any Internet connection. [https://github.com/SensingKit/SensingKit-iOS]
• Carry out a crowdsourcing study and collect data from a number of participants.
• Train lightweight DL models to predict accurately gaze point estimation, using the collected data, and compare against prior baselines/experiments (these should be replicated).
• Carry out ablation studies, personalisation techniques, error analyses, etc.
• Contribute to authoring a scientific article

Eye movement features are considered to be direct signals reflecting human attention distribution with a low cost to obtain, inspiring researchers to augment language models with eye-tracking (ET) data. In this project, we want to investigate how to operationalise eye tracking (ET) features, such as first fixation duration (FFD) and total reading time (TRT), as the cognitive signals to augment LAI models' attention mechanism (e.g., Transformer attention) during training. We want to demonstrate experimentally that augmenting a model with ET data carries linguistic features complementing the information captured by the model and demonstrate that it improves the models' performance. To this end, we will have to collect data from participants and test how different Transformer models benefit from different types of ET attention masks.

The candidate will:

• Carry out controlled studies and collect data from a number of participants.
• Try different approaches to incorporating gaze features in the training process of DL/LLM models and compare against baselines/experiments (these should be replicated).
• Carry out ablation studies, personalisation techniques, error analyses, etc.
• Contribute to authoring a scientific article

Off-line policy learning in RL (from a dataset) is possible with off-policy methods when there is sufficient data. However, the quality of the policy will depend on the quality of the data obtained, and this quality depends (in part) on the amount of data collected. One possibility to improve the data obtained is to apply the knowledge of the symmetry of the system and the environment to do data augmentation of the dataset. In this work we will study:

1. Design of Off-line RL algorithms that take advantage of knowledge of the symmetries of the environment and the system.
2. Combine these systems to make successful initialization of on-policy methods.
3. Try to extend the knowledge of symmetries not by applying data augmentation on the dataset, but by using invariant neural networks.

Seeing that not all neural networks fit to all problems, and that relational data is present in a wide variety of aspects of our daily life, the main focus of this thesis in N3Cat (www.n3cat.upc.edu) and BNN-UPC (www.bnn.upc.edu) is to explore the possibilities of the Graph Neural Networks (GNNs), whose aim is to learn and model graph-structured relational data. We are looking for students willing to study the uses, architectures, and algorithms of GNNs. To this end, the candidate will work on ONE of the following areas:

• Uses: Applying GNNs in emerging application areas, including but not limited to (1) electroencephalogram (EEG) analysis for Alzheimer's disease detection, epilepsy classification, motor imagery; (2) acceleration of the compilation of quantum computing algorithms
• Architectures: Investigating ways to accelerate the processing of GNNs in multiple computing platforms (CPU, GPU, accelerators).
• Algorithms: Developing meta-learning data-driven models to estimate the accuracy of a GNN for a given graph, without training, through synthetic graph generation and correlation analysis.