-
., generation of malicious/destructive code). In this sense, the work includes the study, experimentation, and integration of different security mechanisms applicable to LLMs, as well as the comparison and
-
mechanisms such as access privilege control, guardrails or authorised (or unauthorised) use context, which improve the security alignment of LLMs in the context of their application in psychotherapeutic
-
system and its orchestration, supervision, and validation mechanisms. - Implementing Retrieval Augmented Generation (RAG) mechanisms to enrich the capabilities of the agents.- Developing tools
-
recommendation mechanisms based on semantic analysis and natural language processing, with the aim of facilitating collaboration and convergence of proposals. Developing and training NLP algorithms in multiple
-
the platform. Within the scope of this master's project, only a demonstration of feasibility and proof of concept of implementation is expected, allowing for experimentation and informing the other teams
-
, and interpret the main experimental results in preparation for discussion and thesis writing. 6) Analyse the applicability and adaptation of validated features for anomaly detection in microservice
-
for easy manipulation and experimentation with a system that includes RAG, Tools, and Prompt Engineering, and simulates conversations with real users. The work plan consists of the following development
-
will also be used to improve mechanical performance. Work Plan: The work plan aims at the development of new bio-based curing agents, the use of agro-industrial residues, the production of new