Starting date: October 01, 2022 (flexible)

Application deadline: September 5th, 2022

Interviews (tentative): September 12th, 2022

Salary: 1 975 € gross/month (social security included)

Mission: research oriented (teaching possible but not mandatory)

Keywords: natural language processing, knowledge representation, cultural heritage, transfer learning, multilingualism

CONTEXT

The main challenge of the Patrimalp project is the development of an integrated and interdisciplinary Heritage Science, in order to ensure cultural Heritage sustainability, promotion and dissemination in contemporary society. The ambition is to produce the forms of intelligibility of a global and moving process which starts from the collection of the raw material, its transformation into a primitive object, different lives as a material (alterations, degradations, transformations ...) and as a symbol (relegation, disinterest, oblivion or rebirth, exaltation...) throughout history, and finally from its election as an object of historical and Heritage value and its “promotion” into a work of art. This research is applied to understand how inks and pigments have been conceived over several centuries, how it has been used in art work as well as how the handcrafting method has evolved and been disseminated over centuries and countries.

To make this study possible, the project will gather a large collection of textual material made up of alchemical works and collections of natural or artificial objects collected between the 16th and 18th centuries. To better understand the choice of colors for these "wonders", we want to reconstruct the recipes for making colored material in its context of thought, whether technical or symbolic. These recipes will constitute a new body of research for literary people and a new data-study case for building knowledge about color. This corpus indeed offers modes of representation inscribed in complex forms of writing and fiction whose modalities and frames of reference remain to be analyzed (accounts of technical, medical or physico-chemical experiments inscribed in fictional worlds or mythological, symbolic descriptions of artifacts, or materials collected in nature, mines). On the linguistic level, the inventory of this lexicon in different European and Eastern languages will lead to the formalization of the knowledge of these various skills over time and several cultures. This corpus will thus provide complex data on the material and symbolic origin of the ingredients of color, on its use, its names and its physical or symbolic perception: these data represent a challenge for computer researchers who will have to organize them in order to benefit curators, chemists or physicists, in ontologies representing the state of knowledge from the point of view of scholars over the ages.

To systematically explore the corpus of these recipes, we will use NLP techniques to uncover the correlations between recipes, physical and chemical composition of objects and symbolic references. The final objective is to build a knowledge base (objects, components of objects, materials, colors, know-how, reference framework) each of the parts being able to reference a specific ontology (ontology of pigments, materials, colors...) to make it possible for researchers to observe the trajectory from the writing of color to its technical and artisan practice from this specific corpus.

PHD OBJECTIVES

The PhD project will focus on segmenting, extracting and representing recipes from a corpus of alchemical works from the 16th and 18th centuries to make them accessible to researchers in the humanities. This necessitates to :

- identify which excerpts of the text belong to a recipe; - supervise an annotation campaign to build an analysis and training corpus

- build NLP tools to extract automatically the list of elements (raw material, tools, quantity, units) and actions (verb, adverb, adjective) that made up the recipes;

- analyze the dependencies between the elements of a recipe rules ; - Represent these rules in a formal knowledge representation.

The results of this processing will support : - The documentation of this unique set of text, by inserting the extracted elements to the document meta data to easy retrieval - The building a knowledge base of alchemical recipes

This PhD will need to address several challenges. One of them is to be able to process text composed of multiple non-modern languages (French, German, English, Latin, Greek) [Coavoux2022,Grobol2022] . One approach we will be to study how large multilingual pre-trained models [Delvin2019, Xue2020] can be leveraged and adapted for the task and how disparate collection of corpora of ancient texts can be used to fine-tune them. Another challenge will be the paucity of data for the downstream tasks (segmentation, parsing, Natural Language Understanding [Desot2022]) for this we will need to identify other related corpus (e.g. cooking) to address the problem in a multitask setting (such as NER and NLU) and transfer learning.

SKILLS

- Master 2 in Natural Language Processing, computer science or data science. - Good mastering of Python programming and deep learning frameworks. - Previous experience in text classification, parsing, processing of several languages or text retrieval would be a plus - Very good communication skills in English and good command of French

SCIENTIFIC ENVIRONMENT

The thesis will be conducted within the STEAMER and GETALP teams of the LIG laboratory (http://steamer.imag.fr/ and https://lig-getalp.imag.fr/).The GETALP team has strong expertise and track record in Natural Language Processing, STEAMER team has strong expertise in Knowkledge representation and reasoning.The recruited person will be welcomed within the teams which offer a stimulating, multinational and pleasant working environment. The means to carry out the PhD will be provided both in terms of missions in France and abroad and in terms of equipment (personal computer, access to the LIG GPU servers). The PhD candidate will collaborate with the partners involved in the PATRIMALP project, in particular with Laurence Rivière from the LUHCIE lab (Laboratoire Universitaire Histoire Cultures Italie Europe) and Véronique Adam from the LITT&ARTS lab (Littératures et Arts).

INSTRUCTIONS FOR APPLYING

Applications must contain: CV + letter/message of motivation + master notes + be ready to provide letter(s) of recommendation; and be addressed to Danielle Ziebelin ([email protected]), François Portet ([email protected]) Maximin Coavoux ([email protected])

PhD in ML/NLP – Efficient, Fair, robust and knowledge informed self-supervised learning for speech processing

Starting date: November 1st, 2022 (flexible)

Application deadline: September 5th, 2022

Interviews (tentative): September 19th, 2022

Salary: ~2000€ gross/month (social security included)

Mission: research oriented (teaching possible but not mandatory)

*Keywords:*speech processing, natural language processing, self-supervised learning, knowledge informed learning, Robustness, fairness

*CONTEXT*

The ANR project E-SSL (Efficient Self-Supervised Learning for Inclusive and Innovative Speech Technologies) will start on November 1st 2022. Self-supervised learning (SSL) has recently emerged as one of the most promising artificial intelligence (AI) methods as it becomes now feasible to take advantage of the colossal amounts of existing unlabeled data to significantly improve the performances of various speech processing tasks.

*PROJECT OBJECTIVES*

Recent SSL models for speech such as HuBERT or wav2vec 2.0 have shown an impressive impact on downstream tasks performance. This is mainly due to their ability to benefit from a large amount of data at the cost of a tremendous carbon footprint rather than improving the efficiency of the learning. Another question related to SSL models is their unpredictable results once applied to realistic scenarios which exhibit their lack of robustness. Furthermore, as for any pre-trained models applied in society, it isimportant to be able to measure the bias of such models since they can augment social unfairness.

The goals of this PhD position are threefold: - to design new evaluation metrics for SSL of speech models ; - to develop knowledge-driven SSL algorithms ; - to propose methods for learning robust and unbiased representations.

SSL models are evaluated with downstream task-dependent metrics e.g., word error rate for speech recognition. This couple the evaluation of the universality of SSL representations to a potentially biased and costly fine-tuning that also hides the efficiencyinformation related to the pre-training cost. In practice, we will seek to measure the training efficiency as the ratio between the amount of data, computation and memory needed to observe a certain gain in terms of performance on a metric of interest i.e.,downstream dependent or not. The first step will be to document standard markers that can be used as robust measurements to assess these values robustly at training time. Potential candidates are, for instance, floating point operations for computational intensity, number of neural parameters coupled with precision for storage, online measurement of memory consumption for training and cumulative input sequence length for data.

Most state-of-the-art SSL models for speech rely onmasked prediction e.g. HuBERT and WavLM, or contrastive losses e.g. wav2vec 2.0. Such prevalence in the literature is mostly linked to the size, amount of data and computational resources injected by thecompany producing these models. In fact, vanilla masking approaches and contrastive losses may be identified as uninformed solutions as they do not benefit from in-domain expertise. For instance, it has been demonstrated that blindly masking frames in theinput signal i.e. HuBERT and WavLM results in much worse downstream performance than applying unsupervised phonetic boundaries [Yue2021] to generate informed masks. Recently some studies have demonstrated the superiority of an informed multitask learning strategy carefully selecting self-supervised pretext-tasks with respect to a set of downstream tasks, over the vanilla wav2vec 2.0 contrastive learning loss [Zaiem2022]. In this PhD project, our objective is: 1. continue to develop knowledge-driven SSL algorithms reaching higher efficiency ratios and results at the convergence, data consumption and downstream performance levels; and 2. scale these novel approaches to a point enabling the comparison with current state-of-the-art systems and therefore motivating a paradigm change in SSL for the wider speech community.

Despite remarkable performance on academic benchmarks, SSL powered technologies e.g. speech and speaker recognition, speech synthesis and many others may exhibit highly unpredictable results once applied to realistic scenarios. This can translate into a global accuracy drop due to a lack of robustness to adversarial acoustic conditions, or biased and discriminatory behaviors with respect to different pools of end users. Documenting and facilitating the control of such aspects prior to the deployment of SSL models into the real-life is necessary for the industrial market. To evaluate such aspects, within the project, we will create novel robustness regularization and debasing techniques along two axes: 1. debasing and regularizing speech representations at the SSL level; 2. debasing and regularizing downstream-adapted models (e.g. using a pre-trained model).

To ensure the creation of fair and robust SSL pre-trained models, we propose to act both at the optimization and data levels following some of our previous work on adversarial protected attribute disentanglement and the NLP literature on data sampling and augmentation [Noé2021]. Here, we wish to extend this technique to more complex SSL architectures and more realistic conditions by increasing the disentanglement complexity i.e. the sex attribute studied in [Noé2021] is particularly discriminatory. Then, and to benefit from the expert knowledge induced by the scope of the task of interest, we will build on a recent introduction of task-dependent counterfactual equal odds criteria [Sari2021] to minimize the downstream performance gap observed in between different individuals of certain protected attributes and to maximize the overall accuracy. Following this multi-objective optimization scheme, we will then inject further identified constraints as inspired by previous NLP work [Zhao2017]. Intuitively, constraints are injected so the predictions are calibrated towards a desired distribution i.e. unbiased.

*SKILLS* - Master 2 in Natural Language Processing, Speech Processing, computer science or data science.

- Good mastering of Python programming and deep learning framework. - Previous in Self-Supervised Learning, acoustic modeling or ASR would be a plus

- Very good communication skills in English - Good command of French would be a plus but is not mandatory

*SCIENTIFIC ENVIRONMENT*

The thesis will be conducted within the Getalp teams of the LIG laboratory (_https://lig-getalp.imag.fr/_) and the LIA laboratory (https://lia.univ-avignon.fr/). The GETALP team and the LIA have a strong expertise and track record in Natural Language Processing and speech processing. The recruited person will be welcomed within the teams which offer a stimulating, multinational and pleasant working environment.

The means to carry out the PhD will be providedboth in terms of missions in France and abroad and in terms of equipment. The candidate will have access to the cluster of GPUs of both the LIG and LIA. Furthermore, access to the National supercomputer Jean-Zay will enable to run large scale experiments.

The PhD position will be co-supervised by Mickael Rouvier (LIA, Avignon) and Benjamin Lecouteux and François Portet (Université Grenoble Alpes). Joint meetings are planned on a regular basis and the student is expected to spend time in both places. Moreover, the PhD student will collaborate with several team members involved in the project in particular the two other PhD candidates who will be recruited and the partners from LIA, LIG and Dauphine Université PSL, Paris. Furthermore, the project will involve one of the founders of SpeechBrain, Titouan Parcollet with whom the candidate will interact closely.

*INSTRUCTIONS FOR APPLYING*

Applications must contain: CV + letter/message of motivation + master notes + be ready to provide letter(s) of recommendation; and be addressed to Mickael Rouvier ([email protected]_), Benjamin Lecouteux([email protected]) and François Portet ([email protected]_). We celebrate diversity and are committed to creating an inclusive environment for all employees.

3-year PhD position in Computational Models of Semantic Memory and its Acquisition (Inria and University of Lille, France)

We invite applications for a 3-year PhD position at the University of Lille in the context of the recently funded research project "COMANCHE" (Computational Models of Lexical Meaning and Change). The position is funded by Inria, the French national research institute in Computer Science and Applied Mathematics.

COMANCHE proposes to transfer and adapt neural word embeddings algorithms to model the acquisition and evolution of word meaning, by comparing them with linguistic theories on language acquisition and language evolution. At the intersection between Natural Language Processing, psycholinguistics and historical linguistics, this project intends to validate or revise some of these theories, while also developing computational models that are less data hungry and computationally intensive as they exploit new inductive biases inspired by these disciplines.

The first strand of the project, on which the successful candidate will work, focuses on the development of computational models of semantic memory and its acquisition. Two main research directions will be pursued. On the one hand, we will compare the structural properties associated to different semantic spaces derived from word embedding algorithms to those found in human semantic memory as reflected in behavioral data (such as typicality norms) as well as brain imaging data. The latter data will then used as additional supervision to inject more hierarchical structure into the learned semantic spaces. One the other hand, we intend to experiment with training regimes for word embedding algorithms that are closer to those of humans when they acquire language, controlling the quantity as well as the linguistic complexity of the inputs fed to the learning algorithms through the use of longitudinal and child directed speech corpora (e.g., CHILDES, Colaje). In both cases, both English and French data will be considered.

The successful candidate holds a Master's degree in computational linguistics or computer science or cognitive science and has prior experience in word embedding models. Furthermore, the candidate will provide strong programming skills, expertise in machine learning approaches and is eager to work across languages.

The position is affiliated with the MAGNET team at Inria, Lille [1] as well as with the SCALAB group at University of Lille [2] in an effort to strenghten collaborations between these two groups, and ultimately foster cross-fertilizations between Natural Language Processing and Psycholinguistics.

Applications will be considered until the position is filled. However, you are encouraged to apply early as we shall start processing the applications as and when they are received. Applications, written in English or French, should include a brief cover letter with research interests and vision, a CV (including your contact address, work experience, publications), and contact information for at least 2 referees. Applications (and questions) should be sent to Angèle Brunellière ([email protected]) and Pascal Denis ([email protected]).

The starting date of the position is 1 October 2022 or soon thereafter, for a total of 3 full years.

We have a research engineer position at the University of Montpellier and then at INRAE to coordinate the OntoPortal Alliance, a network in which we develop ontology portals (BioPortal, AgroPortal, EcoPortal, MatPortal, etc.).

This affects semantic web technologies, but not only. We are looking for someone with experience and an interest in project management, open source development projects and international collaborations!

Here is the description: https://bit.ly/3dhX1Ct

You will find below the information for the recruitment of a 24-month IRD post-doc to be filled at the end of 2022 within the UMR Espace-Dev in Montpellier, with missions to be planned in West Africa.

The objective of this post-doc is to develop an IAM (Integrated Assessment Models) approach on the coupling of numerical/mathematical models around zero-emission energy transition issues at the regional level.

Feel free to spread around you,

Contact: benjamin[dot]pillot[at] ird.fr Full information on the position: https://www.espace-dev.fr/news/postdoc-monzemix-ccwamodelling-net-zero-energy-mix-in-climate-changing-western-africa/

kevin chapuis Post-Doc / Research Fellow IRD, UMI 209 UMMISCO / UMR 228 ESPACE-DEV

Mail: [email protected]

In the framework of the European/Japanese e-VITA project (https://www.e-vita.coach/), IMT Atlantique is offering a 15-month post-doctoral position in the field of active living technologies (IoT, data fusion, AI, cloud/edge architectures, user services, coaching, NLP, etc.).

Description and link to apply: https://institutminestelecom.recruitee.com/l/en/o/postdoctorante-ou-postdoctorant-en-fusion-de-donnees-multimodales-cdd-15-mois

The Grenoble Alps University offers a PostDoc position for a highly motivated candidate to be working on the multi-disciplinary research project THERADIA, which aims to create an empathic virtual assistant that accompanies cognitively impaired patients during remediation exercises at home.

The successful candidate will have the exciting opportunity to develop new machine learning techniques for the robust detection of affective and cognitive behaviours from newly collected audiovisual data. Models will be incorporated into the virtual agent to tailor the interaction with the patient, using specific interaction scenarios, and these models will be evaluated and fine-tuned in a clinical trial to demonstrate the effectiveness of the agent in supporting patients suffering from cognitive conditions during digital therapies. If successful, the system will be operated nationally and the cognitive remediation sessions will be covered by social security.

*Duration*: 2 years, *Salary*: according to experience (up to 4142€ / month) *Envisaged starting date*: November 2022

*Scientific environment* The person recruited will be hosted within the GETALP team of the Laboratoire d’Informatique de Grenoble (LIG), which offers a dynamic, international, and stimulating framework for conducting high-level multi-disciplinary research. The GETALP team is housed in a modern building (IMAG) located on a 175-hectare landscaped campus that was ranked as the eighth most beautiful campus in Europe by Times Higher Education magazine in 2018.

*Requirements* The ideal candidate must have a PhD degree and a strong background in machine learning, and affective computing or cognitive science/neuroscience.

The successful candidate should have: - Excellent knowledge of machine learning techniques - Good knowledge of speech and/or image processing - Good knowledge of experimental design and statistics - Strong programming skills in Python - Excellent publication record - Willing to work in multi-disciplinary and international teams - Good communication skills

*Application* Applications are expected to be received on an ongoing basis and the position will be open until filled. Applications should be sent to Fabien Ringeval ([email protected]) and François Portet ([email protected]). The application file should contain:

- Curriculum vitae - Recommendation letter - One-page summary of research background and interests - At least three publications demonstrating expertise in the aforementioned areas - Pre-defence reports and defence minutes; or summary of the thesis with date of defence for those currently in doctoral studies

We invite applications for a 3-year PhD position at the University of Lille in the context of the recently funded research project "COMANCHE" (Computational Models of Lexical Meaning and Change). The position is funded by Inria, the French national research institute in Computer Science and Applied Mathematics.

COMANCHE proposes to transfer and adapt neural word embeddings algorithms to model the acquisition and evolution of word meaning, by comparing them with linguistic theories on language acquisition and language evolution. At the intersection between Natural Language Processing, psycholinguistics and historical linguistics, this project intends to validate or revise some of these theories, while also developing computational models that are less data hungry and computationally intensive as they exploit new inductive biases inspired by these disciplines.

The first strand of the project, on which the successful candidate will work, focuses on the development of computational models of semantic memory and its acquisition. Two main research directions will be pursued. On the one hand, we will compare the structural properties associated to different semantic spaces derived from word embedding algorithms to those found in human semantic memory as reflected in behavioral data (such as typicality norms) as well as brain imaging data. The latter data will then used as additional supervision to inject more hierarchical structure into the learned semantic spaces. One the other hand, we intend to experiment with training regimes for word embedding algorithms that are closer to those of humans when they acquire language, controlling the quantity as well as the linguistic complexity of the inputs fed to the learning algorithms through the use of longitudinal and child directed speech corpora (e.g., CHILDES, Colaje). In both cases, both English and French data will be considered.

The successful candidate holds a Master's degree in computational linguistics or computer science or cognitive science and has prior experience in word embedding models. Furthermore, the candidate will provide strong programming skills, expertise in machine learning approaches and is eager to work across languages.

The position is affiliated with the MAGNET team at Inria, Lille [1] as well as with the SCALAB group at University of Lille [2] in an effort to strenghten collaborations between these two groups, and ultimately foster cross-fertilizations between Natural Language Processing and Psycholinguistics.

Applications will be considered until the position is filled. However, you are encouraged to apply early as we shall start processing the applications as and when they are received. Applications, written in English or French, should include a brief cover letter with research interests and vision, a CV (including your contact address, work experience, publications), and contact information for at least 2 referees. Applications (and questions) should be sent to Angèle Brunellière ([email protected]) and Pascal Denis ([email protected]).

The starting date of the position is 1 October 2022 or soon thereafter, for a total of 3 full years.

Best regards, Angèle Brunellière and Pascal Denis

[1] https://team.inria.fr/magnet/ [2] https://scalab.univ-lille.fr/

The School of Computing and Information Systems at the University of Melbourne is looking for outstanding applicants for two PhD positions to conduct research in Human-Robot Interaction.

About us Home to over 52000 students among which more than 20000 graduates, the University of Melbourne (UniMelb) is one of the world’s top universities (Times Higher Education ranked Melbourne 31st globally; QS: 32nd and ARWU 35th). UniMelb is committed to excellence in research and teaching, interdisciplinary education, and the active promotion of promising young scientists.

The School of Computing and Information Systems (CIS) at Unimelb is a department with a global impact with several research centres among which AI and Digital Transformation. It is located in the new Melbourne Connect building which brings together world-class researchers, industry, SME’s, startups, higher-degree students, artists and Science Gallery Melbourne, in a purpose-built innovation precinct right in the heart of Carlton and next to the Parkville campus. It offers facilities for prototyping with the Telstra Makerspace and for user testing with the UX Lab.

Description We are offering two PhD positions focusing on Human-Robot Interaction. The successful candidate will be able to work in a dynamic group of interdisciplinary researchers and PhD students. These PhD positions will give you access to world class robotic equipment and research facilities. You will work closely with researchers at CIS from the AI and HCI groups in an interdisciplinary environment to share expertise and collaborate on research publications in top conferences and journals in robotics, AI and HCI.

Requirements Please consider applying if you have a Master’s degree or equivalent in: computer science, mechanical engineering, or electrical engineering, or a related field or equivalent practical experience. We especially seek individuals who have knowledge and expertise in the following areas:

Strong programming skills (Python/C++) knowledge and understanding of robot control, machine learning, optimization, and planning Experience with robots, conducting user experiments or AR/VR development Knowing ROS is a greatly appreciated plus

H1 equivalent - First Class Honours (or equivalent) degree Fluency in spoken and written English Good communication skills and ability to present complex content to a diverse audience

High flexibility in acquiring new knowledge Creative and independent thinker Ability to work well in cross-functional and interdisciplinary teams with diverse people

How to apply? Interested applicants should send the following documents via email to [email protected] quoting “PhD Position in Human-Robot Interaction” in the subject line.

A motivation letter (1 page max) describing yourself, your research interests, qualifications, future career goals and research focus and why you would be a suitable candidate A detailed CV

Academic transcripts from your Bachelor’s and Master’s degrees Email addresses of at least two references The position will be filled as soon as possible, and only shortlisted candidates will be notified.

Preferences will be given to applications received before October 7, 2022.

CIS has been pursuing the strategic goal of substantially increasing the diversity of their staff. As an equal opportunity and affirmative action employer, we explicitly encourage nominations of and applications from women as well as from all others who would bring additional diversity dimensions to the university’s research strategies.

The CRIL (Lens Computer Science Research Lab - UMR 8188) Lab at University of Artois (Lens) is hiring a one-year postdoctoral researcher to work on the POSTCRYPTUM project, funded by the Agence Nationale de Recherche (ANR) and AID/DGA. This is a three-years project focused on algebraic cryptanalysis of public key cryptosystems, focused mainly on post-quantum schemes.

For more details on the project, please check here:

https://home.mis.u-picardie.fr/~ionica/postcryptum/Welcome.html

The ideal candidate should hold a Phd degree in Computer Science, Symbolic AI, propositional satisfiability (SAT) and beyond. Skills in one or several of the following topics will be appreciated:

· SAT solving, · Problem encodings and reformulation, · Cryptography, · Pattern mining and machine learning.

The starting date is flexible, but preferably no later than December 2022.

For more information, please contact us ([email protected], [email protected], [email protected]).

We have an open Doctoral position in AI for commercial vehicles at Halmstad University (Sweden). It is offered within a newly granted Swedish (Vinnova-funded) project "Big Data-Powered End User Function Development" with Volvo Trucks (https://www.vinnova.se/en/p/big-data-powered-end-user-function-development-big-fun).

The thesis will seek to apply AI methods to identify moments of potential significance in truck journeys, building a holistic context of truck usage situations and driver journeys.

We would be very grateful if you could please spread it in your networks.

More info about the position and application: https://hh.varbi.com/se/what:job/jobID:533430

Best regards, Fernando Alonso-Fernandez Associate Professor Halmstad University, Sweden https://sites.google.com/view/fernando-alonso-fernandez

Applications are invited for post-doctoral researchers and Ph.D. students in deep learning applied to affective computing (eHealth), video recognition, and medical image processing. Candidates will work at the Laboratory of imaging, vision, and artificial intelligence (LIVIA), ETS Montreal. These funded positions are available immediately and offer a competitive salary. It also offers a possibility for collaborations-internships with top research companies and institutions in Montreal and abroad.

We are looking for highly motivated candidates who are interested in performing cutting-edge research on machine learning algorithms, with a particular focus on deep learning models (e.g, auto-encoders, convolutional and recurrent neural networks) for domain adaptation, information fusion, and weakly-supervised learning. Prospective applicants should have the following profile:

- strong academic record with an outstanding M.Sc./Ph.D. in computer science, applied mathematics, or electrical engineering, preferably with expertise in one or more of the following areas: machine learning, computer vision, pattern recognition, artificial intelligence;

- good mathematical background; - very good knowledge of English

- good programming skills in languages such as Python, with knowledge of deep learning frameworks; - strong publication record in major conferences or journals in computer vision and machine learning.

Application process: For consideration, please send your CV, research statement, names and contact details of two references, transcripts, a link to your Ph.D. thesis, as well as relevant publications to Eric Granger: [email protected]

Eric Granger, ing., Ph.D. Professor and Director of the LIVIA Department of Systems Engineering ETS Montreal, Université du Québec 1-514-396-8650 [email protected]

The ENGAGE (Enabling the next generation of computational physicists and engineers) project invites applications for a PhD Fellowship in 'Deep Learning for Synchrotron X-ray Tomography Data' at the Cyprus Institute, Cyprus.

The PhD project will integrate the strengths brought on by recent advances in Machine and Deep Learning (ML and DL), Computer Vision and Imaging, leveraging novel datasets that will be collected at the BEATS beamline of SESAME (Synchrotron-light for Experimental Science and Applications in the Middle East). It combines opportunities for contribution both in terms of developing novel deep learning architectures, as well as creating a framework for data collection and analysis in the context of 3D Synchrotron CT scans, creating impact across a diverse array of applications ranging from materials science to biology and medicine. In particular, the beamline hosts the first synchrotron full-field X-Ray Computed Tomography (CT) experiment of the region, which will become operational in 2022. A large database of 3D microscopic images will be generated and collected, enabling the study of the internal composition and morphology of materials to scientists from fields including materials science, biology, medicine, and cultural heritage. The topic can be adapted to fit the strengths of the candidate, while more details can be found on the ENGAGE website.

More details: Engage Project 14

Deadline of application: 30 September 2022

How to apply and renumeration: https://engage.cyi.ac.cy/

Candidates must hold a Master's degree or equivalent in computer science, mathematics, engineering, physics or a related field with an outstanding performance record. Experience in machine learning, AI, Signal and Image Processing will also be considered an advantage, while knowledge of deep learning frameworks (e.g. pytorch) will be beneficial. Noting that this is an interdisciplinary topic, and as such excellent candidates that do not cover all related background knowledge will still be considered.

Context Surgical robotics is now widely used with, for instance, more than 5000 Da Vinci systems and one million procedures performed worldwide. Surgery is a complex activity, in a very small anatomical volume, and with a lot of variability between patients and between surgeons. The global objective of the two-year SPARS (Sequential Pattern Analysis in Robotic Surgery: Understanding Surgery) project led by the MediCIS team (LTSI (1), Inserm, Rennes 1 University) is to develop data analysis approaches being able to provide a better understanding of the surgical practice, from complex surgical data. The approaches will be developed thanks to the complementary skills available in the project’s consortium, including time series analysis. In this consortium, the IRISA laboratory (Rennes and Vannes) is calling for applications for a post-doctoral research position (duration two years) on time series analysis.

(1) Laboratoire Traitement du Signal et de l'Image

Missions In the SPARS project context, a trajectory compiles information on the 3D location of the tip of a surgical instrument at the hands of the surgeon, at a constant frequency. The candidate will be mostly involved in one of the three workpackages of the SPARS project. A first task will focus on clustering and classification for such trajectories. Various practical objectives are pursued, including the generation of a model corresponding to a cluster or a class, the characterization of operating modes specific to a type of patient or a type of surgeon, the provision of advice to practitioners in the case of robotic surgeries that are not or not very well documented, the identification of the level of expertise of a practitioner, the prediction of the surgical procedure to be chosen according to the type of patient. These investigations will use dissimilarity measures based on temporal alignment, as DTW [SC71] or elastic kernels as proposed in [CVB07], [CB17] and [M19a]. This task will also address co-clustering for trajectories. The investigations will focus on how to combine time series with other types of data for a co-clustering purpose, using either deep learning [XCZ19] if enough data is available, symbolic representation [BBC15] or latent block [BLN20] models that all need to be adapted to the specificity of kinematics data.

Once a cluster or a class is obtained, another task will be to compute an average trajectory from a set of trajectories. The practical objectives will be the following: highlight deviations from the average trajectory that are potentially interpretable (as characteristics of the practitioner, or of the patient, for example) ; identify the best operating mode to young practitioners or trainees if it is possible to correlate the operating mode with clinical results. Intuitively, on the graphical representation of a time series, variability related to temporality (phase) concerns the abscissa axis, and variability related to shape concerns the ordinate axis. To compute a consensus trajectory, the second task of the package will examine how to extract the atemporal form and the variable component related to temporality, assuming that this atemporal form may be interpreted as an approximation of the consensus. The problem of shape and phase separation has been studied in [PZ16], [SSV10] and [M19a]. The second task will examine how to improve the preliminary work in [M19b], notably by proposing other kernels.

Requirements for this position Doctorate in computer science, applied mathematics and computer science, or mathematics, with a specialization in machine learning and the following requirements:

- theoretical skills and experience in probability / statistics, applied mathematics, machine learning,

- strong knowledge and solid experience in temporal data analysis, - publications in major conferences or journals in the field,

- mastery of data manipulation, relying on machine learning libraries, - programming experience, good programming skills (notably in Python) and technical ability to manage a code development project, - ability to work in a team, and report on the progress of work.

Some knowledge in deep learning will be a plus.

The personal qualities expected are mostly autonomy and interest in interdisciplinarity (health), as well as writing skills (both in French and English). Fluency in French will be a plus.

Work environment Location: Institut de Recherche en informatique et Systèmes Aléatoires (IRISA), Université de Rennes 1 - Campus Beaulieu, 263 Av. Général Leclerc, 35000 Rennes

Duration: 24 months – Applications will be accepted until the position is filled (for recruitment by 1 December 2022 at the latest)

Host team: LINKMEDIA

The successful candidate will work with four academic researchers from IRISA / Rennes / LINKMEDIA team (Simon Malinowski, Associate Professor in Computer Science), IRISA / Vannes / EXPRESSION team (Pierre-François Marteau, Full Professor in Computer Science), LS2N (2) / Nantes / DUKe team (Christine Sinoquet, Associate Professor with French Accreditation to supervise Research (HdR)) and INSERM / Rennes / LTSI MediCIS team (Pierre Jannin, Directeur de recherche INSERM, HdR). The successful candidate will collaborate with the partners in the project, among which the other post-doctoral fellow involved in the project and the project partners experts in surgery and in surgical data analysis.

(2) Laboratoire des Sciences du Numérique de Nantes : UMR CNRS 6004

Income: 2160,26 euros before taxes monthly

How to apply? Documents to be provided :

- detailed Curriculum Vitae including a complete list of publications, - letter of motivation indicating the candidate’s research interests and achievements to date,

- a selection of publications, - the PhD thesis manuscript,

- Master 2 marks (with rank and number of students in the year) - letters of recommendation for the current year,

- contact details of two referees (at least) with whom the candidate has worked (first name, surname, status, institution (give details of acronyms if applicable), city, e-mail address, telephone number)

Questions or application files (zip archive only) should be sent to the four contact persons below:

[email protected] [email protected] [email protected] [email protected] (SPARS project leader)

Simon Malinowksi http://people.irisa.fr/Simon.Malinowski/

Christine Sinoquet https://christinesinoquet.wixsite.com/christinesinoquet

Pierre-François Marteau https://people.irisa.fr/Pierre-Francois.Marteau/

Pierre Jannin https://medicis.univ-rennes1.fr/members/pierre.jannin/index

Please apply here: https://jobs.inria.fr/public/classic/en/offres/2022-05259

Context The team is starting an ambitious research program focused on the teleoperation of humanoid robots (Talos, iCub, Tiago++): an operator wears a mostion capture suit and distant humanoid has to reproduce their movements as precisely as possible.

In this project, our team wants to develop a technology based on predicting the movements of the operator, in particular to compensate for delays and improve the control of the robot.

Missions The main mission of the hired engineer will be to integrate in an open-source library the main algorithms for trajectory prediction. This library will make it possible to evaluate the strengths and weaknesses of the state-of-the-art. It will also give "reference implementations" to the community for trajectory prediction (like stable-baselines for reinforcement learning).

To understand how predictions can be integrated in teleoperation, please read the following article: https://arxiv.org/abs/2107.01281 / video : https://www.youtube.com/watch?v=N3u4ot3aIyQ

The hired engineer will work closely with Jean-Baptiste Mouret (researcher), Serena Ivaldi (researcher), as well as with the PhD students and the post-doctoral researchers of the team.

Environment: Inria Nancy Grand Est, 615 rue du Jardin Botanique, Villers-lès-Nancy (France).

Details and application: https://jobs.inria.fr/public/classic/en/offres/2022-05259

Constructing plans to achieve their goals---and more generally deciding and learning how to act---are fundamental problems for intelligent agents. The focus of cognitive planning is the generation of plans for an individual agent by taking into account other aspects such as reasoning about the beliefs, goals, intentions and actions of other agents ('theory of mind'); learning from past experiences; monitoring the outcome of actions and learning and managing possibly unintended outcomes.

Tools such as logics of knowledge and belief, logics of goals, intentions and actions, game theory and learning should play the role of building blocks of such formalisms. Furthermore, it is assumed that part of the knowledge useful for planning is extracted from a neural network based on spikes which remains to be defined.

The postdoc position is within the ALoRS project (Action, Logical Reasoning and Spiking networks, https://www.irit.fr/ALoRS/) for which IRIT is leader, and is part of workpackage WP2 "Cognitive architectures and NNs", possibly in connection with WP1 "Study and implementation of an SNN". ALoRS (project ANR-21-CE23-0018-01) is a project funded by ANR (the French Research National Agency).

The duration of the contract is 24 months. Gross salary is between 2600 and 3600 Euro, depending on the experience of the candidate. Starting date is in October or November, 2022.

* Context The Institut de Recherche en Informatique de Toulouse (IRIT) is one of the biggest computer science labs in France. Its AI Department has a long-standing tradition of research in knowledge representation and reasoning. IRIT participates in the ANITI project on hybrid AI.

* Profile and Application We are looking for a candidate with a background in knowledge representation and/or machine learning who is interested in their integration in the context of planning et of spiking neural networks. Applicants should propose a brief research project, possibly relating it to our previous work in which we have contributed to cognitive planning in an explicit belief language.

Candidates should send the following to [email protected] and [email protected] before September 21: • CV; • Up to 2 reference letters; • A short proposal of research activities (max 1 page).

Title: Multi-agent trust management for the Internet of Things

Keywords: Trust management systems, Internet of things, Multi-agent system, reinforcement learning

Supervisors: Laurent Vercouter (LITIS, Rouen), Jean-Paul Jamont (LCIS, Valence) Location: Rouen/Valence

Starting date: from October 2022 (it can be a few months later if necessary for the applicant)

Salary: about 1 650 euros per month

Work description for the PhD thesis: When applied to an Internet of things, implementing a trust management system raises new challenges requiring the development of new models and algorithms. Yan et al [2] point out that to have a trustworthy IoT, a trust management system must cover several aspects. It has to allow each entity to evaluate and decide the level of trust for neighboring entities.

The way data is perceived, as well as the way it is transmitted and merged, must also be taken into account. Additionally, trust should be ensured in privacy preservation and during interactions with human users.

At last, an original point is that the identity of entities is not necessarily ensured if the implementation of a robust authentication mechanism is not realistic [3]. This challenges a strong assumption of existing trust management models that use to attach trust values to identities. A new approach for trust management systems in IoT is therefore needed to realistically meet these constraints. The expected work is to define and develop a decentralized multi-level trust management system adapted to the IoT specificities.

The system will have to consider two levels as different targets will be considered: the agent’s own sensors/actuators, and other agents. The work will have a theoretical part in the definition of the models and a practical part for experimentation on the platforms of the MaestrIoT project for which the use of trust management is not yet common [4] and/or has shortcomings [5].

The proposal may be based on the exploitation of work in progress that will be used in the project [6]. One approach considered during the thesis is to develop a reinforcement learning algorithm [7] taking into account the specificities of the internet of things (hardware limitations, low energy consumption, reduced communications, potentially hostile physical environment, ...) to learn trust measures on other agents, sensors and actuators.

This will involve defining a model of desired or feared reference behavior, reward functions based on the observation of the behavior of these entities, an aggregation function to calculate trust values and a decision process that integrates them. An algorithm such as multi-armed bandit would thus allow for the testing of evaluated elements to build a trust model before exploiting it in a decision process.

The algorithms will be experimentally evaluated on the two demonstration platforms of the MaestrIoT project involving mobile robots. Context: The PhD thesis is part of the ANR MaestrIoT (Multi-Agent Trust Decision Process for the Internet of Things) project started in January 2022.

The main objective of the MaestrIoT project is to develop an algorithmic framework for ensuring trust in a multi-agent system handling sensors and actuators of a cyber-physical environment.

Trust management [1] has to be ensured from the perception to decision making and integrating the exchange of information between IoT devices.

The MaestrIoT framework will cover three aspects: (i) definition and recognition of security contexts to evaluate the risks associated to data coming from an agent’s own sensors and from other agents; (ii) definition of a trust management system integrating these security contexts to build and share trust assessments; (iii) sequential decision making processes adapted to information having various trust assessments.

MaestrIoT will consider two privileged application domains: Industry 4.0 and Connected Cooperative Automated Mobility.As part of the ANR MaestrIoT project, we are looking for candidates for a thesis on the development of a trust management system for an Internet of Things, with applications on autonomous vehicle mobile robots and in industry 4.0 for a start of the thesis as soon as possible (October 2002 or shortly after).

A more precise description of the thesis subject is in the attached file.

Interested candidates must send an application file including a CV, transcripts of marks in M1 and M2 (or 4th and 5th year of an equivalent diploma), a cover letter, and letters of recommendation at the latest September 18, 2022 to [email protected] and [email protected]

This Post-doc will be done in the ConfianceAI framework (Confiance.ai). Confiance.ai is the technological pillar of the Grand Défi “Securing, certifying and enhancing the reliability of systems based on artificial intelligence” launched by the French Innovation Council.

It is the largest technological research programme in the #AIforHumanity plan, which is designed to make France one of the leading countries in artificial intelligence (AI).

This post-doc aims to design models to estimate and predict CNN performances without requiring their deployment on the real platform. These models will save time since measurement campaigns will be needed.

Furthermore, it will be possible to test CNN architectures and measure their performances, in terms of execution time, power consumption, memory occupation, etc, before the platforms being available in the market. Thus, measurements such as execution time, or energy consumption can be estimated on an architecture not available from the manufacturer if we know its most "impacting" attributes on the Hardware architecture(s).

This solution makes it possible to quickly explore a large space of configurations (CNN architectures, hardware architectures).

It also makes it possible to avoid developing and optimizing models that will turn out ineffective on the target platform. In this postdoc we will develop new models for CNN performance prediction that respect the rank of the models (Pareto rank-preserving surrogate models).

Hardware-aware Neural Architecture Search (HW-NAS) has recently gained steam by automating the design of efficient DL models.

However, such algorithms require excessive computational resources and thousands of GPU days are needed to evaluate and explore CNN search space.

In this Postdoc we will explore the effectiveness of multi-objective Surrogate Performance Prediction Models in HW-NAS.

Tasks that have to be developed in this postdoc are : - Determination of the CNN search space - Determination of the main characteristics for two of Hardware platforms. This may concern

Qualcomm Hexagon DSPs and edge Nvidia GPUs. - Development and testing of multi-objective surrogate performance prediction models for these 2 types of HW platforms.

Education: A Ph.D in computer/electrical science/engineering is required Salary: 2300 euros/month net (3000 euros/month gross) Deadline for application: 30/09/2022

Duration : 1 year (1 year extension possible) Employer : IRT Systemx

Address: 2 Bd Thomas Gobert, 91120 Palaiseau Start date: Preferably between Sept 1st 2022 and Nov 1th 2022.

An application prepared in English or French should contain:

1. CV with the list of publications. 2. Contact information for 2-3 reference persons. 3. Your most relevant conference or journal publications, in full-text.

For further information contact: - Prof. Smail Niar ([email protected]) LAMIH/CNRS, INSA Hauts-de-France and CNRS, France

- Prof Elghazali Talbi ([email protected]), INRIA and University of Lille, France

Information extraction, Text Recognition in Historical Document Collections LITIS LITIS (Laboratoire d’Informatique, Traitement de l’information et des Systèmes) is a research laboratory associated to the University of Rouen Normandie, Le Havre Normandie Normandie, and School of Engineering INSA Rouen Normandie.

Research at LITIS is organized around 7 research teams which contribute to 3 main application domains: Access to Information, Biomedical Information Processing, Ambient Intelligence. LITIS currently includes 90 faculty staff members, 50 PhD students, 20 PostDoc and Research Engineers. The Machine Learning team of LITIS is developing research in modeling unstructured data (signals, images, text, etc...) with machine learning algorithms and statistical models.

For more than two decades it has contributed to the development of reading systems and document image analysis for various applications such as postal automation, business document exchange, digital libraries, etc... EXO-POPP project: Optical Extraction of Handwritten Named Entities for Marriage Certificates for the Population of Paris (1880–1940)

Thanks to a collaboration between specialists in machine learning and historians, the EXO-POPP project will develop a database of 300,000 marriage certificates from Paris and its suburbs between 1880 and 1940.

These marriage certificates provide a wealth of information about the bride and groom, their parents, and their marriage witnesses, that will be analyzed from a host of new angles made possible by the new dataset. These studies of marriage, divorce, kinship, and social networks covering a span 60 years will also intersect with transversal issues such as gender, class, and origin.

The geolocation of data will provide a rare opportunity to work on places and relocations within the city, and linkage with two other databases will make it possible to follow people from birth to death.

Building such a database by hand would take at least 50,000 hours of work. But, thanks to the recent developments in deep learning and machine learning, it is now possible to build huge databases with automated reading systems including handwriting recognition and natural language understanding. Indeed, because of these recent advances, optical printed named entity recognition (OP-NER) is now performing very well.

On the other hand, while handwriting recognition by machine has become a reality, also thanks to deep learning, optical handwritten named entity recognition (OH-NER) has not received much attention.

OH-NER is expected to achieve promising results on handwritten marriage certificates dating from 1880 to 1923. This project research questions will focus on the best strategies for word disambiguation for handwritten named entity recognition.

We will explore end-to-end deep learning architectures for OH-NER, writer adaptation of the recognition system, and named entity disambiguation by exploiting the French mortality database (INSEE) and the French POPP database.

An additional benefit of this study is that a unique and very large dataset of handwritten material for named entity recognition will be built.

Laboratoire LITIS, EA 4108, Université de Rouen, 76 800 Saint-Etienne du Rouvray, FRANCE Téléphone : (33) 2 32 95 50 13 Fax : (33) 2 32 95 50 22 Email : [email protected]

Missions The research engineer will be in charge of the development of a processing pipeline dedicated to optical printed named entity recognition (OP-NER). He will closely collaborate with a Ph.D. student in charge of Handwritten Named Entity Recognition (OH-NER).

OP-NER is the project’s easiest task and will benefit from the latest results achieved by the LITIS team on similar problems on financial yearbooks. Images are first processed to extract every text information.

This will be achieved with the DAN architecture designed by LITIS which is a deep- learning-based OCR (https://arxiv.org/abs/2203.12273). The research engineer will be in charge of this OCR task.

A benchmark of DAN against available OCR software such as Tesseract and EasyOCR will also be conducted. Then the textual transcriptions will be processed for named entity extraction and recognition.

Named entity recognition is a well-defined task in the natural language processing community. In the EXO-POPP context however, we need to define each entity to be extracted more precisely to make a clear distinction between the different people occurring in the text. For example, we need to distinguish between wife and husband names, and similarly for the parents of the husband and of the wife, and so on for the witnesses, children, etc. An estimation of around 135 categories has been established.

The TAG definition was made by LITIS as well as a first training dataset. Manually tagging the transcriptions has been made possible through the PIVAN web-based collaborative interface (https://litis-exopopp.univ-rouen.fr/collection/12 ).

This platform provides in one single web interface a document image viewer, viewing and editing of OCR results and text tagging facilities for NER. PIVAN eases the annotation efforts of the H&SS trainees and allows for building the large, annotated datasets required for machine learning algorithms to run optimally.

The research engineer will oversee datasets generation and curation as per the requirement of the EXO-POPP NER task, including the handwritten datasets. The named entity recognition task will be based on a state-of-the-art machine learning approach.

We have started some experimentations with the well-known FLAIR NER library (https://github.com/flairNLP/flair). We plan to continue developing and tuning the EXO-POPP named entity recognition module using this library. The research engineer will oversee this task entirely.

Tasks • The research engineer will be in charge of tuning PIVAN for the OCR task. A benchmark of DAN with the available OCR technologies such as Tesseract and EasyOCR will also be conducted.

• The research engineer will be in charge of datasets generation and curation as per the requirement of the EXO-POPP NER task, including the handwritten datasets.

• The research engineer will be in charge of developing the NER module.

Laboratoire LITIS, EA 4108, Université de Rouen, 76 800 Saint-Etienne du Rouvray, FRANCE Téléphone : (33) 2 32 95 50 13 Fax : (33) 2 32 95 50 22 Email : [email protected]

Deliverables: Transcription of the typescript corpus Named entities extracted from the typescript corpus

Skills : • General software development and engineering, Python • Machine Learning, Computer vision, Natural Language Processing • Ability to work in a team, curious and rigorous spirit • Knowledge in web-based programming is a plus

Position to be filled : Positions: 1 Research Engineer Time commitment: Full-time

Duration of the contract: September 1st 2022 – 31st August 2023 Contact: Prof. Thierry Paquet, [email protected]

Indicative salary: €24 000 annual net salary, plus French social security benefits

Location: LITIS, Campus du Madrillet, Faculty of science, Saint Etienne du Rouvray, France