Master's Student - Lisboa, Portugal - Fundação Champalimaud

Fundação Champalimaud
Fundação Champalimaud
Empresa verificada
Lisboa, Portugal

há 4 semanas

João Santos

Postado por:

João Santos

Recrutador de beBee
Descrição

Master's Student at the Natural Intelligence Laboratory*:

Localização:
Centro Champalimaud
Entidade Empregadora: Fundação Champalimaud
Regime: Tempo Inteiro
Posição: Investigação Científica
Início da Candidatura: 15 Maio


2023
Fecho de Candidatura:
05 Jun. 2023A Call for four fellowships for Master's Students is open at Fundação D. Anna de Sommer Champalimaud e Dr.

Carlos Montez Champalimaud (Champalimaud Foundation) in the context of the project entitled "NextGenAI:

Center for Responsible AI", with reference "PRR" - from the call No "crAI 64", funded by the "Agendas Mobilizadoras para a Inovação Empresarial, Plano de Recuperação e Resiliência, Next Generation EU, IAPMEI".

We are a group of neuroscientists and machine learning experts at the Champalimaud Foundation with scientific partners covering a broad range of domain expertises.


Work on the identification of hidden mental states, including emotional and cognitive states, through observable expressions or physiological activity, is crucial.

However, scientific understanding of these phenomena are limited.

These latent states are thought to provide insight into a person's future behaviour, such as what they will say next.

At the same time, Large Language Models (LLMs) have become one of the most exciting frontiers of technological advancement. They provide unprecedented capabilities as tools for everyday language generation tasks. Nevertheless, many unresolved issues must be addressed quickly. For instance, how can we align these generative models with the correct human contexts and objectives to provide maximum benefits to individuals and society? Can we understand the decision-making processes of these complex models to develop more interpretable and explainable human-centered systems? Can they interact with humans in a more positive and fair way such that they remain accessible to non-experts?

We believe that answering these questions will entail the careful identification and interpretation of latent states through advanced methodologies utilising multi-modal data.

You can contribute to this endeavour with us.

Aims of Thesis Project


  • The design and implementation of more objective and interpretable methods for inferring latent states in humans using cost-effective, non-invasive measurement devices. Those include, but are not limited to, electromyography (EMG), functional near-infared spectroscopy (fNIRS), electroencephalography (EEG), speech/audio, eye-tracking, and video and depth cameras (RGB+D).

Experimental design and data collection:

establish an experimental platform and protocol utilizing one or more data modality to reliably record humans while performing decision making tasks.

Methods

General:
Computer Vision, Machine Learning (Deep Learning), Signal Processing, Interpretability in ML

Specific:
Large Language Models (LLMs), Transformers, Pose Estimation, Action Recognition, Prompt Engineering, Virtual Reality, probabilistic models of learning and decision-making, control theory and/or adaptive signal processing

Tools

Python, PyTorch, Scientific Python Stack (Numpy, Pandas, Sci-kit Learn), Data Version Control (DVC), Git, Matplotlib or Other Python-based Visualisation tools, ChatGPT, Unity, Matlab

Data Modalities

Electromyography (EMG), Functional Near-Infrared Spectroscopy (fNIRS), Electroencephalography (EEG), Video and Depth Sensor (Microsoft Kinects or similar), Speech, Virtual Reality

Department

Champalimaud Research Programme | Digital Therapeutics Programme |

Number and type of fellowships or contracts to be awarded

4 fellowship/s.

Application start date

15th May 2023.

Application closing date

5th of June 2023 or until position is filled.

Expected position starting date

Starting date is negotiable; it will depend on the master programme in which the student is enrolled

Required Education Level

Skills/Qualifications

Minimum qualifications


  • Familiarity with some methods, tools, and data modalities from above with measurable experience in key areas (depending on thesis focus);
  • Strong coding skills;
  • A trained scientific mindset;
  • Ability to read and present the latest machine learning, human psychology, or human-computer interaction research to a group of experts;
  • Experimental and task design (depending on thesis focus);
  • Solid experience with data analysis;
  • Excellent level of English.

Benefits:

Full immersion into a research excellence ecosystem with highly motivated researchers, supported by state-of-the-art technology and continuous development opportunities. The monthly remuneration will start at 1000€ month. The fellowship will be paid monthly by bank transfer.

Maximum allowable duration


The Fellowship will have a duration of six months, with the possibility of renewal up to a maximum of 24 months (never surpassing the termination date of the project, expected at 31/12/2025).

Eligibility Criteria

Selection process

Evaluati
Mais empregos da Fundação Champalimaud
Ver todos os empregos de Fundação Champalimaud