Master Thesis

1. Master's Thesis Proposal Form
2. Master's Thesis Inscription Form
3. Master's Thesis Advisor Presentation Authorization
4. Master's Thesis Cover Template

  
Thesis Regulation

 
The Master's Thesis is a mandatory subject of the Master in Artificial Intelligence (MAI), which takes a 30 ECTS effort load.
The Master's Thesis is an academic exercise, where the student will use the acquired knowledge in previous formation courses of the MAI to solve a research problem or a professional problem in the field of Artificial Intelligence. The student must be able to demonstrate her/his skills in the Artificial Intelligence area to solve complex problems. The Thesis content must let to evaluate the strategies and methodologies used to solve the problem, the creativity of the student, the oral and written communication skills to present the work, and the knowledge of the state of the art related to the problem being solved.

The thesis could be a research work, or a technological project. Both kind of works should be done under the supervision of one or more Advisor/s. The Advisor must be a Doctor who is teaching in the MAI. In the case that this is not possible then the advisor must be a Doctor or a Graduate person/s outside the MAI. In the later case, it is needed a MAI Professor acting on behalf of the Advisor ("Ponent"), who will ensure the adequation of the the work developed to the work required fro a Master's Thesis.

Both kind of Master's Thesis (research work and technological project) could be done in an universitary environment or in a company environment.
Master's Thesis developed in a company should be planned within the framework of an Educative Cooperation Agreement ("Conveni de Cooperació Educativa") between the University and the company for each student. This kind of agreements regulate the conditions for the student work at the company, such as the salary, the duration of the time to be devoted to the work, etc.

The steps to finally get a Master's Thesis evaluated by a Committee are:

1. The Master's Thesis theme must be proposed and accepted by the Academic Commission of the MAI (CAIMIA). These proposals could be offered by professors or by people from companies without any student assigned to them, or with a student already assigned to them. The corresponding form is available at Master's Thesis Proposal Form.

2. The Inscription of the Master's Thesis. This step requires a previous acceptance of the Master's Thesis proposal. It is a reservation of the Thesis theme for a determined student, and the compromise of the student to do the work under the supervision of the advisor.  The corresponding form is available at Master's Thesis Inscription Form.

3. Master's thesis Enrolment. This step must be done near the beginning of the term when the student planns to finish the Master Thesis work.

4. Master's Thesis Advisor Presentation Authorization. When the student has finished the Thesis work, he/she must present the Master's Thesis Advisor Presentation Authorization jointly with the Master's Thesis document. The corresponding form is available at Master's Thesis Advisor Presentation Authorization Form.

5. The Master Thesis Document. It must follow the usual scheme of a scientifical/technological work:

 
-  Introduction
- Definition of the problem and definition of the goals
- State of the art/Analysis of related works and tools
- Methodology and strategies to solve the problem
- Technical development (requirements, design, implementation)
- Experimental or Theoretical results
- Evaluation of results
- Discussion (goals/results concordance) and conclusions
- Bibliographical and other references
- Annexes



The cover of the Master's Thesis document must follow the template described in the Master's Thesis Cover Template form.

 

All  Master theses are written and defended in English

2009/2010

1. Anomaly detection in exploratory multivariate time series analysis
2. Interactive visualization of clusters, with application to a medical database
3. Robust visual hierarchical clustering
   
4. Robust feature selection strategies for small samples of spectral data
5. Analysis of EMG information as pseudo-time series
6. Unraveling non-linear dimensionality reduction: A guided tour of high-dimensional data
7. Mapping virtual geographies: data visualization using cartograms
8. Modelling navigation and user’s preferences in a group recommender system
   
9. Learning dependency parsers
10. Query-based sentence relevance ranking
11. Max-margin learning for structured outputs: Applications to NLP Sequential Segmenting Tasks
12. Action Description in Short Term Dyadic Human Interactions
13. Recerca en algorismes d’inducció de regles d’associació i implementació d’un mòdul d’inducció de regles d’associació per a la seva integració en un sistema intel·ligent d’anàlisi de dades
14. Recerca en algorismes d’aprenentatge i raonament basat en casos i implementació d’un mòdul de raonament basat en casos per a la seva integració en un sistema intel·ligent d’anàlisi de dades
15. Recerca en gestió i combinació de coneixement provinent de diverses fonts en una base de coneixement expressada en regles, i la seva integració en un sistema intel·ligent d’anàlisi de dades
16. Recerca en algorismes de clustering amb fluxos de dades continus (data streams) i extensió d’un mòdul de clustering ja existent en un sistema intel·ligent d’anàlisi de dades
17. Implementación de diferentes variantes de Support Vector Machines
18. Selección de atributos con Support Vector Machines
19. Segmentació d’imatges en temps real mitjançant l’anàlisi de textura i color
20. Visualization in parametrizable geometries of sparse data with very high dimension

2009/2010

On-going Theses

1. Business Process Management Specification through Electronic Institutions
2. Algorismes de detecció de comunitats en xarxes complexes
3. Desambiguació semàntica de recursos web retornats per Google
    

Defended Master Theses

2009-2010


Title	Advisor/s	Student	Date	University
			September 10th, 2010	URV
Sublinear Evolutionary Design of Error Correcting Output Codes	Dr. Sergio Escalera (UB), Xavier Baró (UB)	Miguel Ángel Bautista	June 22nd,  2010	UB
Advanced statistical machine learning methods for the analysis of neurophysiologic data with medical application	Dr. Alfredo Vellido (UPC), Dr. Antoni Rodríguez-Frnells (ICREA-UB)	Julià Amengual Roig	June 22nd, 2010	UB
Inferring Best Strategies from the Aggregation of Information from Multiple Agents	Dr. Ulises Cortés (UPC)	Esteve Almirall Mezquita	June 22nd,  2010	UB
			February 16th, 2010	UPC
 
2008-2009


Title	Advisor/s	Student	Date	University
Design and Development of a Case-Based Reasoning Shell integrated in an Intelligent Data Analysis Tool	Dr. Miquel Sànchez-Marrè (UPC)	Beatriz Sevilla Villanueva	September 4th, 2009	URV
			June, 2009	UB
			June, 2009	UB
			June, 2009	UB
			February, 2009	UPC

 

1. A proposal for an Arabic to English SMT System
2. Classification of Biomedical high-resolution micro-CT images for direct volume rendering
3. Desarrollo de algoritmos de elevada eficiencia en computacion cientifica en paralelo
4. Anotació automàtica de recursos Web mitjançant patrons lingüístics
5. Background knowledge to improve induction of decision structures in medicine
6. Anàlisis i comparació d'algorismes pel tractament de fluxes de dades
7. Técnicas de resolución SAT
8. The effect of noise and sample size in the performance of unsupervised feature selection using mixture models
9. Protecció de dades que segueixen sèries temporals
10. Aplicación de Técnicas de IA para mejorar el reconocimiento de promotores a partir de las características físicas de la molécula de ADN.
11. Agent based navigator shared control applied to assistive technologies
12. Joint parsing of syntactic and semantic dependencies

1. Design of a web-accessible agent-based intelligent platform that provides home care services
2. European Home Care Model. Design and implementation of the Actor Profile Ontology
3. How to facilitate the connection between distributed data and agents in the K4Care project