Skip to main content
Search

Zhejiang University

60 related articles

Subscribe to updates
  1. arXiv

    Chemists write only the target, and VibeSyn's agents run a ten-step total synthesis: active human input is 7.3%

    A team led by Zhejiang University built VibeSyn, a hierarchical agent system that turns a chemist's molecular objective into synthesis decisions revised as experimental evidence accumulates, then drives computational tools and automated devices to execute them. In a ten-step catharanthine total synthesis the system recorded 722 actions, of which agent reasoning was 69.7% and tool or device outcomes 18.8%, while active human input and human confirmation were 7.3% and 4.2%; starting from 29.8 mmol of 3-ethylpyridine it produced 160 mg of product. Across three campaigns synthesizing cyclic peptides containing unnatural amino acids, active human input and confirmation were 7.5% and 5.2%.

  2. arXiv

    Agent RL stops scoring whole trajectories: ORSD writes its own rubrics to assign credit token by token

    Online Rubric Self-Distillation (ORSD) has the current policy generate natural-language process rubrics from differences among a group of its own trajectories, lets a frozen verifier pick the bundle whose rubric satisfaction best correlates with task success, and uses that bundle as privileged context: the log-likelihood gap for the same tokens with and without the rubric becomes a token-level correction to the trajectory-level GRPO advantage. On ALFWorld and WebShop, Qwen3-1.7B and Qwen3-4B reach the best overall results among the compared methods under the same environment-interaction budget.

  3. arXiv

    Alibaba DAMO's HoloCT fuses CT reading and 3D segmentation in one model, surpassing all four radiologists in sensitivity and overall accuracy in a multicentre reader study

    Alibaba DAMO Academy's HoloCT is a 3D CT foundation model that puts scan interpretation and voxel-level segmentation inside one vision–language model, so a textual finding can come with a checkable 3D mask. Previously, segmentation networks delineated without interpreting, and CT vision–language models wrote text without voxel-level localization. On 11,455 clinical CT scans from five hospitals, HoloCT-8B raised mean Dice over VoxTell from 0.107 to 0.488 on laryngeal lesions and from 0.543 to 0.715 on lung cancer, and in retrospective reader studies of thyroid tumours and colon lesions it exceeded all four participating radiologists in sensitivity and overall accuracy.

  4. arXiv

    Hyperspectral imaging stops trading highlights for shadows: fusing exposures in the encoded measurement domain lifts spectral dynamic range by 16.0 dB

    Computational hyperspectral imaging must hold highlights and shadows in one shot, but the encoding mask's channels have unequal spectral responsivities, so no single exposure keeps every channel in its reliable range. A Zhejiang University team moves multi-exposure fusion ahead of spectral inversion, into the encoded measurement domain, where an analytical reliability prior and a learned channel-specific adjustment jointly set fusion weights, and measurement-domain spectral-angle supervision constrains relative channel responses. On a 9-channel broadband-encoded camera, an optical-density target gives a spectral reconstruction dynamic range of 42.8 dB, 16.0 dB above the best fixed single exposure at 26.8 dB.

  5. arXiv

    Masked diffusion models lock answers too early: HiLo-RD distills large-window trajectories, lifting GSM8K accuracy by 9.4 points

    Masked diffusion language models can fix answer fragments before the supporting reasoning is resolved when the response window is small, and those fragments then become context for later predictions. HiLo-RD uses the same model's large-window trajectory as a frozen teacher and aligns its hidden representations to cropped small-window views at committed positions. At the same small inference budget, GSM8K accuracy rises from 68.84% to 78.24% while mean readable length falls from 362.8 to 271.3 tokens, and the gains extend to code generation without code-domain training.

  6. arXiv

    Liver CT diagnosis switches to composing findings: diseases unseen in training are detected and localized

    Liver CT diagnosis models usually predict disease labels directly from images, so they only recognize diseases seen in training. DbC splits diagnosis into two steps: it learns a set of radiological findings as dense maps over the scan, then defines each disease by the findings its guidelines require and composes the diagnosis map with a fixed, parameter-free soft-minimum operator. On multi-phase liver CT, for six diseases never seen in training, DbC localized lesions significantly better than every comparator, including models that align images with disease names, and kept its advantage on rare diseases from another institution and on a public external cohort.

  7. bioRxiv

    Super-enhancers no longer need epigenomic profiling: Borzoi-SE calls them from DNA sequence, and CRISPRi confirms HS2 as the β-globin hub

    Super-enhancers could previously be defined only by stitching epigenomic peaks such as H3K27ac, so unprofiled cell states were out of reach and the underlying sequence grammar stayed unread. Borzoi-SE recasts super-enhancer discovery as cell-type-resolved genomic object detection over 524-kb sequences, returning super-enhancer probabilities and boundaries across 90 human cell and tissue lineages. On held-out chromosomes it reached a region-level PR-AUC of 0.643 and F1 of 0.498, against ceilings of about 0.34 and 0.15 for four fragment-based baselines; in silico ablation of the β-globin super-enhancer ranked HS2 as the dominant hub, and CRISPRi silencing of HS2 reduced chromatin accessibility across the locus by 35% to 85%.

  8. arXiv

    History Forcing teaches few-step video models to remember: a 1.3B student reaches 79.82 on 60-second VBench-Long

    Few-step autoregressive video diffusion models gradually lose subject and scene information over long rollouts, and existing distillation targets do not constrain distant history separately. History Forcing builds a context residual from the difference between one autoregressive teacher's predictions with and without extra history, then weights it into the DMD or OPSD distillation target. A 1.3B student reaches a VBench quality score of 85.07 for 15-second generation and a VBench-Long aggregate of 79.82 for 60-second generation, and raises PlayWorld Evolution Quality from 2.35 to 2.70 when the camera leaves and returns.

  9. arXiv

    CellCanon links perturbation prediction, sourced explanation and observation-driven updating in one world state, letting a single new measurement revise unmeasured cell lines

    Single-cell perturbation screens still leave many compounds and cell contexts unmeasured, and CellCanon puts expression prediction, source-grounded explanation and observation-driven updating into one shared world state. A language model compiles drug and target evidence, and numerical readouts then guide single-cell generation; on 339 held-out SciPlex3 conditions it reaches differential-expression AUPRC 0.8067 and Macro-F1 0.8164, against 0.4938 and 0.6961 for the best baseline. In simulated experimental feedback, one measured source context per drug lifts differential-expression AUPRC to 0.9077 for unmeasured cell lines with a fixed-weight updater.

  10. aclanthology.org

    Black-box LLM backdoors now show up in reasoning graphs: GCoTD lifts TPR by 3.44 points on GPT-4o

    GCoTD turns a black-box LLM's reasoning into a graph and uses a lightweight detector to judge whether a response was triggered by a backdoor, without weights, gradients or training data. Earlier detectors leaned on white-box signals, labeled poisoned data or output-level signatures, which are brittle on closed-source APIs. Across four attack types and six models, GCoTD reaches the highest or near-highest TPR in most settings, improving TPR by 3.44 and 3.75 percentage points over the prior state of the art on GPT-4o and LLaMA-3-8B respectively.