Sort by
Refine Your Search
-
Category
-
Program
-
its diagnostic capabilities. While pre-trained models will be included, we will continue to explore feature-based methods. Feature-based methods offer a level of explainability necessary for end-users
-
the weighting factor given to each parameter. In this process abstentions are not allowed. Candidates shall be ordered by applying the successive voting method. The jury has the faculty to not approve a candidate
-
: OBJECTIVES | FUNCTIONS - Develop a forecast framework using federated learning - Evaluate the impact of FL in the EVs management - Help in the implementation of FL methods - Assist the research team in
-
automated reasoning methods, enhanced by and integrated with the capabilities of LLMs to extract relevant information, for aligning the PA specifications with the guidelines. BINDING LEGISLATION Law 40/2004
-
the page level. To address this limitation, the student will integrate the scraper generation techniques developed in the first task of the project with a new static analysis method for approximating
-
to the development of an implementation setting for alignment methods for complex PA processes interacting with data objects. These methods should take inspiration from algorithmic techniques developed in data-aware
-
retrieval, specifically looking at (a) distillation of large retrieval models into small and efficient dual encoders, and (b) training of learned sparse retrieval methods. BINDING LEGISLATION Law 40/2004
-
the candidates according to their classification consisting on the sum of the partial classifications assigned in each evaluation criterion, and considering the weighting factor given to each parameter. In
-
according to their classification consisting on the sum of the partial classifications assigned in each evaluation criterion, and considering the weighting factor given to each parameter. In this process
-
to their classification consisting on the sum of the partial classifications assigned in each evaluation criterion, and considering the weighting factor given to each parameter. In this process abstentions are not allowed