Deadline for receipt of applications: February 14th, 2025 23:59 AoE (15 February 2025, 13:59h Europe/Madrid Time)
IMDEA Networks’ Pervasive Wireless System Group headed by Prof. Domenico Giustiniano invites applications for PhD positions in the framework of HORIZON-MSCA-2023-DN-01-01 ANT Project
Host institution: IMDEA/SFI
Country: Spain
Supervisor: Prof. D. Giustiniano [IMDEA] / D. Frometa (SFI)
Co-supervisors: Prof. J. Yang [TU Delft]; Prof. D. Giustiniano [IMDEA]; Dr. D. Pau [ST]
Objectives:
1) To design a data curation pipeline that collects farming data of different modalities and different quality (e.g., depending on local IoT sensor processors)
2) To develop a data integration pipeline that integrates farming data into highly energy-efficient 1-bit foundation models (BitFM) that can be deployed on edge devices in the farm, with a decision mechanism for the activation of data integration pipeline taking into account the network situation, the amount of harvested energy, and the frequency of tasks .
3) To design cost-efficient strategies to improve reliability of trained BitFM by augmenting its responses with output from proprietary FMs (e.g. GPT-4) when necessary.
Expected Results:
1) Definition of suitable emerging Agritech use cases to apply 1-bit foundation models
2) System architecture of data curation and integration into BitFM for smart farming, and evaluation in different network and energy conditions
3) Methods for fine-tuning low-confidence responses of BitFM by integrating proprietary FMs, tested in realistic deployments in smart agriculture.
PhD enrolment: Doctoral School of UC3M
Planned secondments:
Candidate profile: telecommunication engineering, applied mathematics, electrical engineering, computer science (in order of preference)
Desirable skills/interests: signal processing, statistical filtering, machine learning, embedded systems, applied optimization (the applicant should be proficient in at least one or two of the skills)
Application Deadline: February 14, 2025, AoE
Submit Your Application HERE
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No. 101169439.