Northwestern University

Directly imaging exoplanets requires subtracting the star's glare from the image, and standard practice builds that model from reference-star images, so telescope pointing errors and thermal drifts between reference and science observations limit sensitivity. FLIP implements a physics-based instrument model in PyTorch and uses automatic differentiation to optimize millions of parameters in stages, pre-calibrating on a reference star and then tuning on the science image to compensate those instabilities. On JWST/NIRCam coronagraph data, FLIP achieves 2 to 15 times better sensitivity than the standard KLIP method and makes the first detection of 51 Eridani b at 4.3 micrometers in F430M at signal-to-noise 6.3.

Directly imaging exoplanets requires subtracting the star's glare from the image, and standard practice builds that model from reference-star images, so telescope pointing errors and thermal drifts between reference and science observations limit sensitivity. FLIP implements a physics-based instrument model in PyTorch and uses automatic differentiation to optimize millions of parameters in stages, pre-calibrating on a reference star and then tuning on the science image to compensate those instabilities. On JWST/NIRCam coronagraph data, FLIP achieves 2 to 15 times better sensitivity than the standard KLIP method and makes the first detection of 51 Eridani b at 4.3 micrometers in F430M at signal-to-noise 6.3.

A prospective multicenter diagnostic study tested whether an AI system could read standardized "blind sweep" ultrasound images and estimate gestational age in real clinics, on mobile devices, with novice operators and unfamiliar hardware. The model was minimally tuned on a small pilot set of 120 participants in Chicago, with Nairobi site data held out as a strictly independent external test; in a primary cohort of 385 participants between 16 and 36 weeks, AI gestational-age estimates were comparable in accuracy to the current standard of care from expert operators using traditional cart-based ultrasound.

A prospective multicenter diagnostic study tested whether an AI system could read standardized "blind sweep" ultrasound images and estimate gestational age in real clinics, on mobile devices, with novice operators and unfamiliar hardware. The model was minimally tuned on a small pilot set of 120 participants in Chicago, with Nairobi site data held out as a strictly independent external test; in a primary cohort of 385 participants between 16 and 36 weeks, AI gestational-age estimates were comparable in accuracy to the current standard of care from expert operators using traditional cart-based ultrasound.

TCR–pMHC binding prediction no longer has to start from a docked complex. Structure-based predictors co-fold the pair and read binding off the interface, so placement error changes which residues and contacts reach the model. EMBER folds each molecule as an unbound monomer, encodes both as rigid-motion-invariant residue graphs, and compares them through bidirectional attention over CDR3 and peptide residues. Across four folding tools and five epitope-disjoint folds it leads every metric, with 26% to 85% higher cross-tool ranking agreement than structure-based baselines.

TCR–pMHC binding prediction no longer has to start from a docked complex. Structure-based predictors co-fold the pair and read binding off the interface, so placement error changes which residues and contacts reach the model. EMBER folds each molecule as an unbound monomer, encodes both as rigid-motion-invariant residue graphs, and compares them through bidirectional attention over CDR3 and peptide residues. Across four folding tools and five epitope-disjoint folds it leads every metric, with 26% to 85% higher cross-tool ranking agreement than structure-based baselines.

An agent's performance depends not only on its underlying model but on its runtime harness: prompts, memory, skills, tools and orchestration. Harness2 makes harness improvement a two-level recursion. At the task level, a contrastive propose-probe-compose step compares execution behavior before and after edits without ground-truth answers; at the harnessing level, that process is distilled into an experience library that guides later tasks. On WorkBuddy-Bench, JobBench and Harvey's Legal Agent Benchmark, a single step improved both GLM-5 and Gemini 3.5 Flash across four model-harness configurations, by up to 11.5 points, 4.5 points and 11.8 percentage points.

An agent's performance depends not only on its underlying model but on its runtime harness: prompts, memory, skills, tools and orchestration. Harness2 makes harness improvement a two-level recursion. At the task level, a contrastive propose-probe-compose step compares execution behavior before and after edits without ground-truth answers; at the harnessing level, that process is distilled into an experience library that guides later tasks. On WorkBuddy-Bench, JobBench and Harvey's Legal Agent Benchmark, a single step improved both GLM-5 and Gemini 3.5 Flash across four model-harness configurations, by up to 11.5 points, 4.5 points and 11.8 percentage points.

The authors introduce SpatialRNN, in which recurrent neuron-to-neuron communication is replaced by a spatially evolving medium governed by discretized PDEs, and prove that the medium state equals a convolution over the entire history of hidden states, i.e. a structured infinite-order RNN with a fixed parameter count; constructive conditions confine gradient modes to circles of prescribed radii, yielding better long-horizon performance than other recurrent models with fewer parameters.

The authors introduce SpatialRNN, in which recurrent neuron-to-neuron communication is replaced by a spatially evolving medium governed by discretized PDEs, and prove that the medium state equals a convolution over the entire history of hidden states, i.e. a structured infinite-order RNN with a fixed parameter count; constructive conditions confine gradient modes to circles of prescribed radii, yielding better long-horizon performance than other recurrent models with fewer parameters.

The work proposes STAVE, which replaces in-context demonstrations with two task-specific vectors added to the input embeddings of a frozen LLM or LMM, a readout vector updating the answer-producing tokens and a context vector updating the other structural token groups, both trained with answer labels on prompts with and without demos; across six LMMs and five LLMs, STAVE matches or outperforms per-layer interventions with far fewer task parameters, surpasses 15-shot ICL and prior task vectors on 18 text tasks, and runs at zero-shot inference cost.

The work proposes STAVE, which replaces in-context demonstrations with two task-specific vectors added to the input embeddings of a frozen LLM or LMM, a readout vector updating the answer-producing tokens and a context vector updating the other structural token groups, both trained with answer labels on prompts with and without demos; across six LMMs and five LLMs, STAVE matches or outperforms per-layer interventions with far fewer task parameters, surpasses 15-shot ICL and prior task vectors on 18 text tasks, and runs at zero-shot inference cost.

The work presents RDTU, a Residual Diffusion framework for time-series unlearning: it first uses a retained-set neural tangent kernel predictor to obtain a deletion-compatible base forecast, then quantifies the global and local structural support of each affected window via the volume contribution of the retained-reference data, and finally has a diffusion model generate a residual correction that estimates the counterfactual forecast, yielding a pseudo-label field that guides a lightweight model update; experiments show RDTU consistently produces unlearned models that most closely match exact retraining.

The work presents RDTU, a Residual Diffusion framework for time-series unlearning: it first uses a retained-set neural tangent kernel predictor to obtain a deletion-compatible base forecast, then quantifies the global and local structural support of each affected window via the volume contribution of the retained-reference data, and finally has a diffusion model generate a residual correction that estimates the counterfactual forecast, yielding a pseudo-label field that guides a lightweight model update; experiments show RDTU consistently produces unlearned models that most closely match exact retraining.

The work introduces SRMA-Mamba, a Mamba-based network that uses the Spatial Anatomy-Based Mamba (SABMamba) module to perform selective scans across sagittal, coronal, and axial anatomical planes for a global spatial context, and the Spatial Reverse Mamba Attention (SRMA) module to progressively refine boundaries from a coarse segmentation map and hierarchical encoder features, reporting better segmentation metrics than SegResNet, UX-Net, MedNeXt, SwinUNETR, SwinUNETRv2, and SegMamba on CirrMRI600+ T1W and T2W, with fewer parameters and GFLOPs than SegMamba.

The work introduces SRMA-Mamba, a Mamba-based network that uses the Spatial Anatomy-Based Mamba (SABMamba) module to perform selective scans across sagittal, coronal, and axial anatomical planes for a global spatial context, and the Spatial Reverse Mamba Attention (SRMA) module to progressively refine boundaries from a coarse segmentation map and hierarchical encoder features, reporting better segmentation metrics than SegResNet, UX-Net, MedNeXt, SwinUNETR, SwinUNETRv2, and SegMamba on CirrMRI600+ T1W and T2W, with fewer parameters and GFLOPs than SegMamba.