Skip to main content
Search

Institut Polytechnique de Paris

7 related articles

Subscribe to updates
  1. arXiv

    LungMend fills the cavities severe-TB CT masks miss, using only a healthy-lung shape prior and reaching 0.986 Dice

    On severe tuberculosis CT, cavities and consolidations have soft-tissue density and are left out of lung masks by segmenters such as R231. LungMend treats mask correction as shape completion: a latent diffusion prior trained on 57 healthy lungs, conditioned on a synthetically corrupted mask and on bone anatomy that disease does not alter, with no pathological data or CT intensities. On 200 radiologist-corrected slices from 20 severe TB cases it reaches 0.986 Dice and 0.57 mm average surface distance, with cavity and consolidation recall of 0.993 and 0.981.

  2. arXiv

    6G slice placement stops relying on one fixed plan: ORION feeds verifier-checked outcomes back to its LLM planner

    Placing, routing and admitting multi-domain 6G network slices is an NP-hard online problem, fixed planning strategies reverse rank across topologies, and most rejected requests do have a feasible placement. ORION's hierarchical orchestrator uses two LLM agents to translate service intents and generate abstract placement plans, and records each attempt's network conditions, plan shape and verifier feasibility outcome in an episodic memory that returns to the planner as labelled examples. With selection and execution held fixed on held-out topology families, the memory-equipped planner admits more than both the memory-off and first-in-first-out variants on two seeds, though the magnitude is not established.

  3. arXiv

    Constrained decoding is not the reasoning killer: with matched prompts, validity constraints gain up to 17.4%

    Constrained decoding is often assumed to degrade LLM reasoning, but once the prompt and expected output format are aligned, causal ablations show the apparent gap comes mainly from prompt-level differences rather than decoder masking. The work splits constraints into format, validity and reasoning types and tests them across four model families and eleven open-source models; validity constraints, which remove invalid answers, give the most consistent gains, up to 17.4%, while reasoning constraints help only when the enforced trajectory is itself beneficial.

  4. arXiv

    A network first differentiates, then recoheres: after the data-constraint graph connects, the learned state keeps reorganizing

    On an 11-bit parity task, the Boolean function state learned by a Transformer first differentiates and then recoheres as data accumulate, while the connectivity of the training-set constraint graph rises monotonically, showing that the learned state keeps reorganizing after the identifiable support stops changing.

  5. arXiv

    BAC pairs identity-linked boxes with LoRA tuning to reach 93.20% person-centric QA at 8B scale

    Using LSMDC v2 movie clips, the work builds an automatic character-identification and identity-linked spatial grounding pipeline, manually verifies a benchmark of 750 clips and 3,000 person-centric questions, and compares five identity-grounding representations; visual face boxes combined with textual coordinates prove most consistent across scales, and LoRA fine-tuning of Qwen yields BAC, with BAC-8B reaching 93.20% overall person-centric QA accuracy.

  6. medRxiv

    Tumour region identification guided scoring (TRIGS) and foundation model-based Tumour Infiltrating Lymphocyte scoring are prognostic for pathological complete response/event free survival in the triple negative patients in the PARTNER randomized controlled trial

    This study updated an automated tumour infiltrating lymphocyte (TIL) assessment pipeline, proposing TRIGS aligned with clinical scoring guidelines and foundation model-based SAM-TIL, and showed in 166 neoadjuvantly treated patients in the TransNEO cohort that they predicted pathological complete response with odds ratios of 1.95 (95% CI 1.22-3.03, p=0.005) and 2.32 (95% CI 1.43-3.77, p=0.001), and in 277 triple negative and HER2-positive patients in The Cancer Genome Atlas showed overall survival hazard ratios of 0.79 (95% CI 0.63-1.00, p=0.05) and 0.80 (95% CI 0.67-0.97, p=0.02), with correlation to gold standard clinical assessment of 0.59-0.69 and no substantial difference from gold standard assessment in predicting pathological complete response (AUC 0.60-0.

  7. arXiv

    Geometric Self-supervised Pretraining on 3D Protein Structures Using Subgraphs

    This work proposes a new self-supervised pretraining task for 3D graph neural networks: pretrained on 542k SwissProt protein structures from the AlphaFold Database, the model predicts the Euclidean distance between the geometric centroid of each protein subgraph (2-hop ego networks centered on 10% of amino acids) and the global geometric centroid of the whole protein, discretized into 10 equal bins and trained with cross-entropy; across ProNet, SchNet, and GCN backbones and ca_base, ca_angles, and ca_bb featurizations, it improves Fold, Superfamily, Family, and React classification by up to about 6% over no-pretraining and edge-distance pretraining baselines, without multiple views, augmentations, or masking strategies.

You’re all caught up