Two PhD Student Positions (Research Assistants) on Value Alignment for Large Language Models W/M

Description

Professor Andrea Cavallaro is looking for two PhD students (Research Assistants) in the field of value alignment for Large Language Models (LLMs).

The increasingly natural formats of interaction between human and machines helps overcoming barriers for users. However, these advances come with the risk of potential harms emerging from such interactions. Content generated using LLMs is becoming widely available and progressively less distinguishable from content generated by humans. Human interaction with systems using LLMs and exposure to content generated by LLM-based tools raise several questions and may entail paradigm shift in terms of personalization of information access. Meaningful personalization and access to trusted information sources is only possible if such systems embed our values and adapt to human’s needs and preferences.

The first PhD project will investigate the design and development of usable value alignment approaches for LLM companions to interact with users with different age and level of literacy. The second PhD project will investigate how to enable an LLM to suitably manifest itself to humans and explain its outputs to different individuals, user groups or communities. A particular care will be devoted to verifying the existence of any biases in the generated outputs and to systematically test the aligned LLM to improve its robustness against misalignment.

This initial brief outlines the research topics. We will work collaboratively with the chosen PhD candidates to refine the research questions and develop a PhD plan. The successful candidates will also become a doctoral student at EPFL ( http://www.epfl.ch ) conditional on parallel application to, and acceptance by, the EPFL Doctoral School ( http://phd.epfl.ch/applicants ), EDEE Program. Appointment for a Research Assistant / PhD position is for a maximum of 4 years, subject to successful progress, and should lead to a PhD dissertation. Starting date can be negotiated. Queries related to these positions can be sent to Sarah Delporte ( sarah.delporte@idiap.ch ).

Required language

Fluency in written and spoken English. 

About Idiap

Idiap is a research institute of national importance that engages in fundamental research, education and technology transfer in artificial intelligence, machine learning and signal processing. Idiap offers competitive salaries and conditions in a young, high-quality, dynamic, and multicultural environment. Idiap is located in the town of Martigny in Valais, a scenic region in the south of Switzerland, surrounded by the highest mountains of Europe, and offering exceptional quality of life, exciting recreational activities, including hiking, climbing and skiing, as well as varied cultural activities. English is the official working language.

At Idiap we place great emphasis on diversity and we know diversity fosters creativity and innovation. We are committed to equality of opportunity, to being fair and inclusive, and to being a place where we all belong. Employment at Idiap is based solely on a person's merit and qualifications. Idiap does not discriminate against any employee or applicant because of race, color, religion, gender, sexual orientation, gender identity/expression, national origin, disability, age, marital status, pregnancy, or any other basis protected by law.

For frequently asked questions (FAQs) about living in Switzerland, please go to https://www.idiap.ch/en/faq

Required languages

• English - Level advanced

Réf: fabfb90f-482c-4e6b-82cb-1a50818dc3b7