Division of labor and task specialization are key elements explaining the remarkable ecological success of human and animal societies. Social insect colonies are characterized by a highly effective division of labor, with workers specializing in brood care early in life and in foraging later in life. Various theoretical models have been proposed to explain division of labor, most prominently the response threshold model, which postulates that individuals differ in their response threshold to task-associated stimuli, and will engage in particular tasks depending on this threshold. While this model has received some experimental support, the current data still do not explain how division of labor is implemented for all different tasks of a social insect colony. How is division of labor implemented? The success of social insect colonies lies in the capacity of all its members to behave in a well-organized manner, which involves elaborate communication among colony members. Accordingly, ants, wasps and bees use a wide range of pheromones, intraspecific chemosensory messages, to regulate almost every aspect of their life. Astonishingly, the role of olfaction, the main sensory modality used by insects, on the division of labor has been greatly overlooked. Using the honey bee Apis mellifera as a model, I propose and will investigate the groundbreaking hypothesis that variation in olfactory perception and processing, especially regarding social pheromones, can give rise to division of labor. I will use a novel approach in social insects, based on newly developed neurogenetic tools. I will compare olfactory perception and processing in the hive’s different worker task groups (nurses, foragers etc.) and will experimentally manipulate the activity of neural circuits involved in pheromone information processing, while following bees’ behavior and task allocation. In the end, Olf@Task aims to definitively establish the role of olfactory perception on the division of labor.

To understand the history of the chondritic asteroid Ryugu and whether it had a role in delivering volatiles and organic material to early Earth, SDR will constrain the mineral and geochemical variability of sediment samples from Ryugu on a grain and sub-grain scale. Because Ryugu is a rubble-pile asteroid made of re-accreted pieces of the parent body/bodies, near-surface sediment collected by the Hayabusa2 mission represents many depths within the interior of the parent. However, current analyses of collected reflectance spectra are done at the bulk sample scale which can average over interesting features to determine the major components. As individual grains are prepared to be sent out for proposed science, each grain is being documented from multiple viewing angles using the MicrOmega hyperspectral imager (0.99-3.65 micrometers) developed at IAS. This has created an enormous, rich dataset of ~20 hyperspectral images for each of 350 grains analyzed so far, and this dataset will continue to grow over the coming months. By conducting novel statistical analyses of this dataset of >7000 high-resolution image cubes (22 micrometers/pixel), I will identify co-occurring mineral phases, minor phases, and variation in organic phases that have implications for the aqueous alteration conditions within Ryugu and its parent body. Statistical analyses will include: principal component analysis, k-means clustering, distinctiveness measures, pixel-to-pixel correlation, edge detection, and network centrality. My experience publishing on statistical analyses of existing, noisy datasets has prepared me to extract meaningful information from this large, complicated dataset. This work will culminate in an updated model of formation and alteration on Ryugus parent body which relies on the types and position of minerals to infer alteration gradients and constrain the nature, number, and temperature of aqueous alteration events.

Book publishing involves many stakeholders and a complex set of inflexible tools and formats. Current workflows are inefficient because authors and editors cannot make changes directly to the content once it has been laid out. They are costly because creating different output formats, such as PDF, ePub or HTML requires manual labor. Finally, new requirements such as the European directive on accessibility incur additional costs and delays. The goal of the OnePub POC project is to demonstrate the feasibility and value of a book production workflow based on a set of collaborative editing tools and a single document source representing the “ground truth” of the book content and layout. The editing tools will be tailored to the needs of each stakeholder, e.g. author, editor or typesetter, and will feature innovative interaction techniques from the PI’s ERC project ONE. The project will focus on textbooks and academic publications as its testbed because they feature some of the most stringent constraints in terms of content types and content layout. They also run on tight deadlines, emphasizing the need for an efficient process. OnePub will define a unified format for the document source, create several collaborative document editors, and develop an open and extensible architecture so that new editors and add-ons can be added to the workflow. Together, these developments will set the stage for a new ecosystem for the publishing industry, with a level-playing field where software companies can provide components while publishers and their service contractors can decide which components to use for their workflows.

Artificial intelligence is a pervasive and ubiquitous technology, with fast developments and ever-growing applicative horizons. While achieving outstanding results, estimating the decisions’ reliability is a critical and open problem with strategic outcomes: from estimating uncertainty in numerical simulations to harnessing system reactions in open environments. Uncertainties both in input data and model output should be handled to increase the confidence in AI applications : physics-based models, edge computing, data frugal approaches, interactions with humans, with many applications in energy, climate change adaptation, bioinformatics, engineering, fundamental and material science,… DeMythif.IA is an international doctoral training and career development program driven by UPSaclay, and its 19 research and industrial partners supported by thematic networks to enhance the scientific excellence and career development of 30 PhD fellows in Greater Paris area. The program focuses on 3 interdisciplinary scientific axes 1) handling uncertainties on data and model and quantifying uncertainties on the predictions, 2) managing explainability, so that the trained model outcome can be trusted and explained to a human, and 3) encouraging frugality, both in term of labelled data and energy for training, so that real-life applications (as opposed to proofs-of-concept) emerge. The fellows will benefit from world-class scientific programs, quality supervision and industrial secondments and also from personalized career development activities and transferable skills training (interpersonal, communication, digital, entrepreneurship, open science, gender, ethics). DeMythif.AI’s strength lies in its diversity permitting to fine-tune the content offered to the fellows from a professional perspective. It ultimately allows concretizing dynamic research collaborations and developing strong synergies around the academic and industrial community, opening new perspectives beyond the project.

Last year’s discovery of states hosting the fractional quantum Hall effect at zero magnetic field in a lattice system – zero-field fractional Chern insulators (FCIs) – has caused tremendous excitement within the condensed matter community. These zero-field FCIs were first observed – by a team including the applicant – using both optics and transport in a molybdenum ditelluride (MoTe2) moiré superlattice. These states, analogs of the Jain sequence FQH states, host Abelian anyon excitations. However, for certain twist angles, moiré MoTe2 is also predicted to host states with non-Abelian anyons - the main requirement for achieving scalable topologically protected quantum computation. These non-Abelian states have yet to be observed. Perhaps the largest difficulty in moiré systems is sample quality and twist angle control. Contact resistance, strain, twist angle inhomogeneity, bubbles, and polymer residue can preclude the formation and observation of delicate non-Abelian FCIs. The main goal of TOPOMAX is to investigate, using optical and transport measurements, the existence of non-Abelian states at zero magnetic field in ultra-high quality twisted MoTe2 heterostructures. To achieve this, we will implement cutting-edge fabrication techniques that enable unprecedented twist angle control and homogeneity while allowing for devices compatible with optical and transport measurements. In particular, the bent bilayer nanomanipulator technique will enable systematic investigation of zero-field FCIs as a continuous function of twist angle. Determining whether the moiré MoTe2 system hosts non-Abelian states would lay the groundwork for study of zero-field non-Abelian anyons – and perhaps, one day, topologically protected qubits.