HAL-MINES ParisTech
Not a member yet
27348 research outputs found
Sort by
Managing generativity in innovation: from control to emergence: Call for Paper
No full textInternational audienceEmerging technologies such as generative AI or autonomous systems introduce new forms of generativity, namely the capacity to produce unexpected, evolving, and selfpropagating outcomes. These technologies challenge conventional management approaches. They do so by producing outcomes that cannot be fully anticipated or predefined, thereby undermining the assumptions of stability, linearity and alignment theorized in traditional management and innovation literature. Importantly, the emergence of AI provides new challenges on, for instance, how can organizations manage generative technologies? What practices, processes and structures enable organisations to leverage generative technologies? This track explores how organisations understand, design, and manage generativity as a new condition for creativity and transformation. We invite theoretical, empirical, and design-oriented contributions examining how generativity reshapes organisational creativity, learning, and strategy. The track aims to bring together researchers studying the creative, organisational, and epistemic implications of generative technologies. Description:Digitalisation has transformed the way organisations innovate, yet a new phase is unfolding, one that is characterised not merely by automation or connectivity, but by generativity. Generative technologies, including GenAI, synthetic biology, computational design, and adaptive systems, create artefacts, knowledge, and possibilities that evolve autonomously and perpetually. These technologies are not only tools or enablers; they are complex systems, capable of evolving autonomously and reconfiguring the very environments in which they operate (Thomas & Tee, 2022). This shift challenges traditional managerial models founded on prediction, control, and eRiciency. Managing innovation in generative contexts requires new forms of creativity, coordination, and learning (Haefner et al., 2021).</div
Efficient Fine-Tuning of DINOv3 Pretrained on Natural Images for Atypical Mitotic Figure Classification (MIDOG 2025 Task 2 Winner)
No full textAtypical mitotic figures (AMFs) represent abnormal cell division associated with poor prognosis. Yet their detection remains difficult due to low prevalence, subtle morphology, and inter-observer variability. The MIDOG 2025 challenge introduces a benchmark for AMF classification across multiple domains. In this work, we fine-tuned the recently published DINOv3-H+ vision transformer, pretrained on natural images, using low-rank adaptation (LoRA), training only ~1.3M parameters in combination with extensive augmentation and a domain-weighted Focal Loss to handle domain heterogeneity. Despite the domain gap, our fine-tuned DINOv3 transfers effectively to histopathology, reaching first place on the final test set. These results highlight the advantages of DINOv3 pretraining and underline the efficiency and robustness of our fine-tuning strategy, yielding state-of-the-art results for the atypical mitosis classification challenge in MIDOG 2025
Platform competition and strategic trade‐offs for complementors: Heterogeneous reactions to the entry of a new platform
No full textInternational audienceAbstract Research Summary We study how the entry of a rival platform affects the strategies of the incumbent's complementors. The latter face a trade‐off: While the entry threatens their benefits from indirect network effects, it also allows them to escape intense within‐platform competition. Studying Epic Games' entry into the PC video game market—until then dominated by Steam—we show that this trade‐off does not resolve uniformly, driving heterogeneity in strategic reactions. Complementors with weaker strategic resources (independent developers) were more likely to multihome and became less responsive to the incumbent's attempts to orchestrate collective action through platform‐wide sales promotions. In contrast, complementors more reliant on indirect network effects (multiplayer developers) were less likely to multihome and became more responsive to orchestration attempts. Managerial Summary As competition between digital platforms intensifies, complementors—firms that provide complementary products or services—must adjust how they engage with platform owners. The entry of a rival platform creates both opportunities and risks: it offers an alternative with less intense competition but also fragments the user base that underpins network benefits. We find that these opposing effects shape complementors' behavior differently. Those with fewer strategic resources are more likely to join the new platform and become less willing to follow the incumbent's coordinated initiatives. In contrast, complementors whose products rely strongly on network effects tend to remain with the incumbent and cooperate more closely with its orchestration efforts. For managers, this highlights that platform competition not only shifts market dynamics but also reshapes the motivations and strategies of heterogeneous complementors
J-NeuS: Joint field optimization for Neural Surface reconstruction in urban scenes with limited image overlap
No full textInternational audienceReconstructing the surrounding surface geometry from recorded driving sequences poses a significant challenge due to the limited image overlap and complex topology of urban environments. SoTA neural implicit surface reconstruction methods often struggle in such setting, either failing due to small vision overlap or exhibiting suboptimal performance in accurately reconstructing both the surface and fine structures. To address these limitations, we introduce J-NeuS, a novel hybrid implicit surface reconstruction method for large driving sequences with outward facing camera poses. J-NeuS leverages cross-representation uncertainty estimation to tackle ambiguous geometry caused by limited observations. Our method performs joint optimization of two radiance fields in addition to guided sampling achieving accurate reconstruction of large areas along with fine structures in complex urban scenarios. Extensive evaluation on major driving datasets demonstrates the superiority of our approach in reconstructing large driving sequences with limited image overlap, outperforming concurrent SoTA methods
Open Data from LIGO, Virgo, and KAGRA through the First Part of the Fourth Observing Run
No full textInternational audienceLIGO, Virgo, and KAGRA form a network of gravitational-wave observatories. Data and analysis results from this network are made publicly available through the Gravitational Wave Open Science Center. This paper describes open data from this network, including the addition of data from the first part of the fourth observing run (O4a) and selected periods from the preceding engineering run, collected from May 2023 to January 2024. The public data set includes calibrated strain time series for each instrument, data from additional channels used for noise subtraction and detector characterization, and analysis data products from version 4.0 of the Gravitational-Wave Transient Catalog
Insight into molecular grafting of carbon aerogels for electrochemical capacitors applications
No full textInternational audienceMolecular grafting of carbon aerogels, made from the polycondensation of resorcinol and formaldehyde, was studied in order to enhance their specific capacitance in electrochemical capacitors. Grafting of redox entities, anthraquinone, was performed via aryl diazonium reduction. While aryl diazonium salt to carbon aerogel molar ratio does not seem to have an impact on the number of chemisorbed or physisorbed anthraquinone, increasing reaction time, even up to 24 h, leads to an increase of chemisorbed anthraquinone moieties. Physisorption of anthraquinone occurs at the initial stage of the modification and does not evolve with time or molar ratio. Interestingly, physisorbed and chemisorbed anthraquinone moieties are all electroactive and despite an inevitable decrease of the specific surface area, an important increase of the specific capacity of anthraquinone functionalized carbon aerogel up to 84 mAh.g−1 is observed
Conditional Generative Models for High-Resolution Range Profiles: Capturing Geometry-Driven Trends in a Large-Scale Maritime Dataset
No full textHigh-resolution range profiles (HRRPs) enable fast onboard processing for radar automatic target recognition, but their strong sensitivity to acquisition conditions limits robustness across operational scenarios. Conditional HRRP generation can mitigate this issue, yet prior studies are constrained by small, highly specific datasets. We study HRRP synthesis on a largescale maritime database representative of coastal surveillance variability. Our analysis indicates that the fundamental scenario drivers are geometric: ship dimensions and the desired aspect angle. Conditioning on these variables, we train generative models and show that the synthesized signatures reproduce the expected line-of-sight geometric trend observed in real data. These results highlight the central role of acquisition geometry for robust HRRP generation.</div
GWTC-4.0: Population Properties of Merging Compact Binaries
No full textInternational audienceWe detail the population properties of merging compact objects using 158 mergers from the cumulative Gravitational-Wave Transient Catalog 4.0, which includes three types of binary mergers: binary neutron star, neutron star--black hole binary, and binary black hole mergers. We resolve multiple over- and under-densities in the black hole mass distribution: features persist at primary masses of and with a possible third feature at . These are departures from an otherwise power-law-like continuum that steepens above . Binary black holes with primary masses near are more likely to have less massive secondaries, with a mass ratio distribution peaking at , potentially a signature of stable mass transfer during binary evolution. Black hole spins are inferred to be non-extremal, with 90% of black holes having , and preferentially aligned with binary orbits, implying many merging binaries form in isolation. However, we find a significant fraction, 0.24-0.42, of binaries have negative effective inspiral spins, suggesting many could be formed dynamically in gas-free environments. We find evidence for correlation between effective inspiral spin and mass ratio, though it is unclear if this is driven by variation in the mode of the distribution or the width. (Abridged
Understanding Deep Tech: Conceptual Foundations, Strategic Stakes, and Policy Implications
No full textInternational audienceDeep tech has become a central concept in innovation, entrepreneurship, and public policy, frequently invoked to describe science-based ventures and breakthrough technologies. Despite its growing prominence, the term remains conceptually ambiguous and theoretically underdeveloped. This article adopts a conceptual–analytical approach to clarify the meaning and role of deep tech within contemporary innovation systems. Drawing on innovation studies, technology entrepreneurship, and science and technology studies, it examines how deep tech is defined, mobilized, and operationalized across technological, business, and policy domains. The analysis shows that deep tech is best understood not as a specific technology category, but as a performative and strategic innovation regime characterized by radical innovation, long time horizons, ecosystem dependence, and mission-oriented public intervention. The article offers conceptual clarity and practical insights for managers and policymakers navigating high-uncertainty, science-based innovation contexts
Eco-friendly Conductive biopolymer nanocomposites and Life Cycle Assessment: a review
No full textInternational audienceConductive bionanocomposites are attracting growing interest as multifunctional materials. They can meet the requirements of electrical applications while supporting sustainable development. This review summarizes recent research on bionanocomposites made from biopolymer matrices and carbon conductive fillers that can be processed by additive manufacturing. These materials offer several advantages, including reduced dependence on fossil resources, possibility of low-impact processing, minimized risks in case of dissemination, and satisfactory electrical properties with low amounts of conductive fillers. However, despite their “green' label, their actual environmental performance has not been fully demonstrated. Only a limited number of comprehensive Life Cycle Assessments (LCA) are available. This review discusses the potential of these materials, while underscoring the necessity for rigorous environmental analysis. Such assessments are essential to validate their sustainability from a circular economy perspective using LCA