19/11/2020
The Human Resources Strategy for Researchers
Marie Skłodowska-Curie Actions

PhD position 19 – MSCA COFUND, AI4theSciences (PSL, France) - “Learning dynamics in biological and artificial neural networks”

This job offer has expired


  • ORGANISATION/COMPANY
    Université PSL
  • RESEARCH FIELD
    NeurosciencesNeuroinformatics
  • RESEARCHER PROFILE
    First Stage Researcher (R1)
  • APPLICATION DEADLINE
    26/02/2021 23:00 - Europe/Brussels
  • LOCATION
    France › Paris
  • TYPE OF CONTRACT
    Temporary
  • JOB STATUS
    Full-time
  • HOURS PER WEEK
    35
  • OFFER STARTING DATE
    01/09/2021
  • EU RESEARCH FRAMEWORK PROGRAMME
    H2020 / Marie Skłodowska-Curie Actions COFUND
  • REFERENCE NUMBER
    AI4theSciences-PhD-19
  • MARIE CURIE GRANT AGREEMENT NUMBER
    945304

OFFER DESCRIPTION

“Artificial intelligence for the Sciences” (AI4theSciences) is an innovative, interdisciplinary and intersectoral PhD programme, led by Université Paris Sciences et Lettres (PSL) and co-funded by the European Commission. Supported by the European innovation and research programme Horizon 2020-Marie Sklodowska-Curie Actions, AI4theSciences is uniquely shaped to train a new generation of researchers at the highest academic level in their main discipline (Physics, Engineering, Biology, Human and Social Sciences) and master the latest technologies in Artificial Intelligence and Machine Learning which apply in their own field.

26 doctoral students will join the PSL university's doctoral schools in 2 academic cohorts to carry out work on subjects suggested and defined by PSL's scientific community. The 2020 call will offer up to 15 PhD positions on 24 PhD research projects. The candidates will be recruited through HR processes of high standard, based on transparency, equal opportunities and excellence.

 

Description of the PhD subject: “Learning dynamics in biological and artificial neural networks”

 

Context - Motivation

Category learning in biological networks

Learning is  a fundamental property of the cortex that enables a broad range of cognitive functions. Amongst them, speech acquisition forces our brain to create new internal auditory categories for speech-relevant sounds, and to associate new stimuli to preexisting categories. However,​ how category learning shapes neural representations in the brain is an open question.

In the human brain, neural populations selective to speech acoustics have been identified in ​non-primary regions of the human auditory cortex (Norman-Haigneré et al. 2018). Ferret electrophysiology shows that learning to categorize acoustic patterns alters responses in multiple auditory cortical regions (Atiani et al. 2014), with prominent and magnified representations in ​non-primary auditory areas (Elgueda et al. 2019). These results suggest that neural circuits for categorizing sounds are found in higher auditory areas. But how these specialized circuits are shaped during learning remains unknown.

Category learning in artificial networks

This ability to​ categorize diverse signals into different meanings and classes is also a key ability of Deep Neural Networks (DNNs). How DNNs transform complex and noisy signals into useful abstractions is the subject of intensive investigations in computer science, and has already revealed important connections with neuroscience. Specifically, both the cortex and DNNs share central computational principles such as the notion of hierarchical organizations and of spatio-temporal convolutions (Hassabis et al. 2017). These shared elements make​ artificial networks a potentially useful model for understanding hierarchical effects of categorization in sensory cortex​ (LeCun et al. 2015).

Stimulus representations in sensory cortical areas and DNNs exhibit many non-trivial similarities. Specifically, several studies have demonstrated that DNNs are able to linearly predict neural responses in visual (Yamins & DiCarlo. 2016), auditory (Kell et al. 2018) and language responsive brain regions (Caucheteux & King. bioRxiv 2020). These results suggest that supervised training on visual and auditory tasks leads ​trained DNNs to ultimately generate representations similar to those of the brain.

Learning dynamics in neural networks

DNNs are thus being proposed as models of sensory and language processing in the brain. However, whether the learning dynamics in these models are similar to those in the brain is unknown​. This issue is critical both for building better models of the brain and for building machines that match the learning capacities of human listeners.

The overarching hypothesis of this project is that​ computational constraints force artificial and biological neural networks to adopt similar learning dynamics​. We will test this hypothesis by explicitly training ferrets and DNNs on a speech-related auditory task. We will then monitor (i) how DNNs distribute their learning resources​ during training (​Goals 1&2​), (ii) where and when changes occur in auditory cortex ​during ​training (​Goal 2​), and (iii) whether functional changes induced by training are similar between the two systems (Goal 3​).

The temporal sequence of such changes is, to date, unknown, partly because it is technically challenging to monitor individual neurons along the course of learning. To tackle this question, our laboratory will perform state-of-the-art neuroimaging using high-resolution neuroimaging in the ferret (Bimbard et al. 2018; Landemard et al. biorXiv 2020) to track learning dynamics in primary and non-primary fields of auditory cortex. Only a few studies have investigated how stimulus representations in DNNs change during learning, but preliminary evidence suggests that lower layers of DNNs stabilize much earlier than representations in higher layers (Raghu et al. 2017). Here, we will​ compare learning dynamics in (1) the auditory cortex and (2) DNNs throughout their respective training on the same task​. We postulate that training should induce prominent changes in speech-evoked responses of later (non-primary) stages, as found in humans. Once the learning dynamics are characterized in both cortical and artificial networks, we will ​investigate single unit activations in DNNs through all learning stages​. Co-supervisors Yves Boubenec and Jean-Rémi King are experts in UltraSound imaging and DNN-based neuroscientific analyses, respectively, as demonstrated by their recent publications (Bimbard et al. 2018, Landemard et al. bioRxiv 2020, Caucheteux & King. bioRxiv 2020).

 

Scientific Objectives, Methodology & Expected results

  • Goal 1:​ developing an artificial model of auditory cortex through exposure to natural sounds​​ (M1-M6) The first goal of the project will be to formulate a theoretical framework with a DNN trained to categorize environmental sounds into broad categories. We will compare neural and simulated representations of test stimuli with Representational Similarity Analysis (RSA; Kriegeskorte et al. 2008). We will extend previous human findings to ferrets showing that first layers of DNNs correlate more with representations in the primary auditory cortex, whereas deep layers correlate more with those of higher auditory regions. This project is highly​ interdisciplinary as it involves advanced machine learning with quantitative analysis of neuroimaging big data. Neural data for Goal 1 are already available in the group of co-supervisor Yves Boubenec.
  • Goal 2: identifying the changes of representations in DNN and auditory cortex after training ​​(M7-M18) We shall train ferrets to categorize speech from other sounds, behaviorally signified by a Two-Alternative Forced Choice (2AFC) paradigm. Analogous to the animal training, we will re-train the DNN on the same task as ferrets with arbitrarily defined category boundaries between speech and other sounds. Our prediction is that post-training representation of sounds becomes less (or more) distant if they belong to the same (or different) categories. Based on our previous work on the cortical encoding of speech acoustics (Norman-Haigneré et al. 2015, Landemard et al. bioRxiv 2020), we expect these differential representations to emerge in later (non-primary) stages of the processing hierarchy. Changes in categorical representations will be modeled by adding back a new classification layer to the last embeddings and performing supervized fine tuning of the whole network with the training stimuli used in the animal.
  • A 2-month ​international secondment with Sam Norman-Haigneré (Columbia University) is scheduled during this period. He is an expert in the encoding of speech in the human brain (Norman-Haigneré et al. 2015, 2018), as well as a collaborator of co-supervisor Yves Boubenec for cross-species comparison of natural sound encoding in auditory cortex.
  • Goal 3: tracking similarities in learning dynamics across artificial and biological neural networks ​​(M19-M30) We will seek common learning dynamics shared between biological and artificial neural networks. For this purpose, we will predict cortical activity from DNN activation along the course of learning. We will deploy dimensionality reduction techniques such as Canonical Correlation Analysis (CCA) to capture shared learning dynamics in subspaces relevant for both auditory cortex and DNN. Common neural dynamics will be dissected by studying DNN single units activation, which is not feasible in the auditory cortex. This goal will benefit from the ​intersectorial background of co-supervisor Jean-Rémi King, who is a research scientist both at the École Normale Supérieure-PSL and Facebook Artificial Intelligence Research.
  • The last 6 months (M31-36)​ of the PhD will be dedicated to finishing paper redaction and writing the thesis dissertation.

This project addresses fundamental problems in neuroscience that link sensory representations of complex sounds to their abstract meanings, which touches upon their utilization in symbolic systems such as music and language.

 

International mobility

Sam Norman-Haigneré (MIT) will be a close collaborator of this project. He is an expert in auditory neuroscience and functional neuroimaging, themes that are central to this proposal and has experience coupling machine learning with the analysis of functional neuroimaging data (Kell et al. 2018). The student will move for a period of 2 months to the USA in order to benefit from Norman-Haigneré’s advice and expertise.

 

Thesis supervision

Yves Boubenec and Jean-Rémi King

 

PSL

Created in 2012, Université PSL is aiming at developing interdisciplinary training programmes and science projects of excellence within its members. Its 140 laboratories and 2,900 researchers carry out high-level disciplinary research, both fundamental and applied, fostering a strong interdisciplinary approach. The scope of Université PSL covers all areas of knowledge and creation (Sciences, Humanities and Social Science, Engineering, the Arts). Its eleven component schools gather 17,000 students and have won more than 200 ERC. PSL has been ranked 36th in the 2020 Shanghai ranking (ARWU).

 

More Information

Benefits

  • Opportunity to conduct academic research in a top 100 university in the world.
  • High-quality doctoral training rewarded by a PhD degree, prepared within Ecole Normale Supérieure - PSL and delivered by PSL.
  • Access to cutting-edge infrastructures for research & innovation.
  • Appointment for a period of 36 months (job contract delivered by the involved component school of PSL) based on a salary of 3100 € gross employer (including employer tax) per month or approximately a 2228 € gross salary per month.
  • Job contract under the French labour legislation in force, respecting health and safety, and social security: 35 hours per week contract, 25 days of annual leave per year (“congés annuels”). Eventual complementary activities may be accepted or proposed by the co-supervisors (maximum of 64h/year for teaching, 32 day/year for specific missions).
  • Short stay(s) or secondment in France or abroad are expected.
  • An international environment supported by the adherence to the European Charter & Code.
  • Access to AI training package, with a strong interdisciplinary focus, together with a Career development Plan.

Eligibility criteria

  • Applicants must have a Master’s degree (or be in the process of obtaining one) or have a University degree equivalent to a European Master’s (5-year duration) to be eligible at the time of the deadline of the relative call.
  • There is no nationality or age criteria, but applicants must not have resided or carried out their main activity (work, studies, etc.) in France for more than 12 months in the 3 years immediately before the deadline of the call (MSCA Mobility rule).
  • Applicants must declare to be available to start the programme on schedule.

For submitting your online application, go to: https://www.psl.eu/recherche/grands-projets-de-recherche/projets-europee...

 

The online application should contain the following documents:

  • English translated transcripts from the Master’s degree (or equivalent 5-year degree). A copy of the Master’s degree or a certificate of achievement will be required later on for the final registration.
  • International curriculum vitae and a cover letter explaining the reasons that lead him/her to prepare a PhD, why he/she applies to this offer and his/her professional project (guidelines will be given to the applicants in order to help him/her in the writing of his/her letter).
  • Two academic reference letters.
  • A statement duly signed on the mobility rules, availability, and conflicts of interest.

 

The applicants can only apply to one PhD project among the available ones. Multiple applications of one candidate will automatically make all his/her applications ineligible.

Selection process

The applications will be analysed by the Management Team for eligibility and completeness. Afterwards, the applications will be reviewed by the Selection Committee. In the pre-selection round (March-April 2021), applicants will be rated using a scoring system based on 3 criteria (academic excellence, experience, motivation, and qualities). A shortlist of qualified applicants will be interviewed during the selection round (June 2021) to further assess their qualifications and skills according to the predefined selection criteria.

All information regarding the applications (criteria, composition of the Selection Committee, requirements) can be found on the website of the programme, in greater detail.

 

The selection and recruitment processes of the PhD student will be in accordance with the European Charter for Researchers and Code of Conduct of the Recruitment of Researchers. The recruitment process will be open, transparent, impartial, equitable, and merit based. There will be no discrimination based on race, gender, sexual orientation, religion of belief, disability, or age.

Additional comments

The LSP (Laboratory of Perceptive Systems, UMR 8248) is an audio-visual research unit in the Department of Cognitive Science at the École Normale Supérieure (ENS - PSL). The LSP has 9 PIs, 6 from audition, 3 from vision, and combines theoretical and experimental research, both audio-visual psychophysics and behavioral animal experiments including chronic recordings from the cortex.

LSP is a perfect environment for this project, as it combines expertise in audition (one of the lab professor, Shihab Shamma, is a world-recognized expertise in auditory perception and auditory physiology) and in machine learning.

 

The Ecole Normale Supérieure - PSL is a leading multidisciplinary institution that focuses on training through research. The ENS - PSL defines and applies scientific and technological research policies, from a multidisciplinary and international perspective and counts close relationships with prestigious partners, in France and abroad. It encompasses fourteen teaching and research departments, spanning the main humanities, sciences, and disciplines. The ENS - PSL currently has a staff of almost 800 lecturers, ENS - PSL, CNRS or associated researchers, and post-doc researchers. Within its Departments, the ENS - PSL includes 40 research units identified as ENS - PSL, INSERM or INRIA, encompassing ENS - PSL and CNRS agents as well as 300 foreign researchers and 650 doctoral students. The ENS - PSL respects the principles of the European Charter and Code for Researchers and is engaged in the HRS4R certification.

Web site for additional job details

Required Research Experiences

  • RESEARCH FIELD
    Mathematics
  • YEARS OF RESEARCH EXPERIENCE
    1 - 4

Offer Requirements

  • REQUIRED EDUCATION LEVEL
    Mathematics: Master Degree or equivalent
  • REQUIRED LANGUAGES
    ENGLISH: Excellent

Skills/Qualifications

  • Strong mathematical and/or computational background.
  • An interest in neuroscience is more than recommended since the goal of the project is to combine machine-learning techniques with analysis of large neuroimaging datasets to address fundamental questions in the field of neuroscience and machine learning.

Work location(s)
1 position(s) available at
LSP, Ecole Normale Supérieure - PSL
France
Paris
75005
29, rue d Ulm

EURAXESS offer ID: 579017

Disclaimer:

The responsibility for the jobs published on this website, including the job description, lies entirely with the publishing institutions. The application is handled uniquely by the employer, who is also fully responsible for the recruitment and selection processes.

 

Please contact support@euraxess.org if you wish to download all jobs in XML.