arXiv subjects

Wen Li

Publications and source records attributed to Wen Li.

At least 19 recordsLinked to original sources

arXiv2026

DWISE: Denoising Workload-Informed Scheduling Engine

Unlike autoregressive large language models (AR-LLMs), diffusion LLMs (dLLMs) generate text by iteratively denoising many tokens in parallel, offering a promising path to efficient inference. Existing serving frameworks, however, schedule dLLM requests with policies designed for AR models, namely first-come-first-served (FCFS) admission and prefill priority. We find that both policies are ill-suited to dLLMs, wasting much of their speed. FCFS ignores that a dLLM request's decoding cost is highly heterogeneous and unknown at arrival, causing severe head-of-line blocking. Prefill priority, widely used in AR-LLM serving to keep time-to-first-token low, sharply cuts dLLM efficiency, as rounds that run only prompts leave most of the batch idle. We present DWISE, a Denoising Workload-Informed Scheduling Engine for efficient dLLM serving. DWISE schedules requests by their remaining work, measured in a block-level unit shared by prefill and denoising, so both phases share a round. To predict the unknown denoising workload, DWISE combines a lightweight prompt-based predictor at admission with a single refinement from the first denoising round's hidden states, which capture model-specific information absent from the prompt and require no additional model execution. Implemented in SGLang and evaluated on five models from three dLLM families under bursty workloads, DWISE reduces mean end-to-end latency over SGLang by 5.4$\times$ on average and up to 14.9$\times$ while preserving task accuracy. The code will be released publicly.

cs.LG↗

arXiv2026

TensorBound: Characterizing Tensor Binding Failures in AI Compiler and Runtime Stacks

AI compiler and runtime stacks must preserve bindings between tensor values and their operations, storage, and execution state. These bindings can fail while the underlying objects remain present, with symptoms appearing in another layer. We present TensorBound, an empirical study of tensor binding failures in IREE and the XLA and PyTorch ecosystems. The study analyzes 264 reports and reconstructs 149 distinct repaired causes through source tracing and comparisons of pre-fix and repaired revisions. Each reconstruction connects the first invalid decision to its violated condition, physical effect, and repair. Missing eligibility checks, incorrect mappings or translations, and lost or omitted relations account for 69.1% of repaired causes. These patterns highlight the need to preserve binding facts and check the conditions for their use. The 101 first invalid decisions in compiler code identify candidate checking points before execution. Four archived IREE diagnostic attempts distinguish effect localization from source-based cause explanation. Together, these findings motivate transition-specific checks that preserve binding facts across representations and enforce the required conditions at consuming decisions.

cs.CR↗

arXiv2026

Evaluating a Large Language Model for Radiotherapy Patient Scheduling under Realistic Clinical and Operational Constraints: a Feasibility Study

Radiotherapy scheduling must balance clinical constraints and operational objectives. We evaluated the feasibility of using a large language model (LLM) to generate treatment schedules through simulation and retrospective clinical comparison. A one-year simulation modeled 1,400 new patients across 12 treatment categories and three linear accelerators (LINACs). Structured prompts provided ChatGPT with patient information, schedule status, clinical rules, and operational objectives. Scenarios assessed treatment-time consistency, LINAC continuity, gap reduction, and handling of infeasible requests. LLM schedules were compared with heuristic and constraint programming (CP-SAT) approaches under two objective settings. A six-week retrospective analysis included 2,043 clinical appointments. LLM-generated schedules met all predefined feasibility constraints in the evaluated scenarios, with 98.6-98.8% of fractions within preferred treatment-time windows. Prioritizing LINAC continuity reduced switching from 52.4 +/- 2.6% to 9.4 +/- 0.6%. Gap-constrained temporal relaxation reduced average daily gaps to 47.1 +/- 8.5 minutes and Friday gaps to 89.2 +/- 18.4 minutes while maintaining low overtime. The framework identified predefined violations in infeasible requests and suggested corrective actions. Heuristic and CP-SAT methods also generated feasible schedules, with performance dependent on objective weighting. In retrospective evaluation, all approaches scheduled appointments without overtime; LLM strategies reduced mean gaps by 50.8-77.1% and improved treatment-time consistency from 86.7% to 99.5-99.9%. These findings support further evaluation of structured LLM scheduling as a human-in-the-loop decision-support approach for radiotherapy.

physics.med-ph↗

arXiv2026

The Energetic Particle Detector Instrument on ELFIN

We present the design methodology, implementation, and data products of the Energetic Particle Detector (EPD) instrument on-board the Electron Losses and Fields Investigation (ELFIN) dual SmallSat mission. ELFIN was designed to study the time-evolution of energetic particle distributions in Earth's magnetosphere, having operated in polar Low Earth Orbit (LEO) from 2018 to 2022. Each satellite's instrument payload contained a pair of detector heads to sense differential-directional fluxes of 50 keV to $\sim$5 MeV electrons (EPDE), and 50 keV to $\sim$2 MeV atomic ions (EPDI), with an energy resolution better than 50% ($ΔE/E<0.5$). The satellites were spinning with a nominal period of 2.8 s in a plane containing Earth's background magnetic field, with particle counting intervals subdivided into 16 consecutive spin sectors. Each EPD was paired with a three-axis Fluxgate Magnetometer (FGM) sampled at 80 Hz for local pitch-angle determination. Each detector head operated on the principle of pulse-height analysis (PHA) in solid state media, collimated by a rectangular 22.5$^\circ$ field of view and entrance aperture. The EPDE contained a stack of 6 silicon detectors (0.7 cm total thickness) with 0.15 cm$^2$ str geometric factor, with ion contamination mitigated by an aperture foil and digital coincidence logic. The EPDI contained a stack of 2 silicon detectors (0.1 cm total thickness) with 0.03 cm$^2$ str geometric factor, with electron contamination mitigated by a magnetic broom stage and digital anti-coincidence logic. The total assembled mass, volume, and power required for the EPD detector heads and circuit boards were 939 g, 10$\times$9.8$\times$8.4 cm$^3$, and 2.3 W, respectively. We describe each EPD design element alongside instrument data products and processing.

physics.ins-det↗

arXiv2026

AegisAsync: Symbolic Async-Event Simulation for Risks in Python Asynchronous Programs

Security and robustness risks in Python async programs depend on event order, object identity, and execution after exceptions or cancellation. Detecting these risks requires connecting related operations across function calls and tracking their effects along execution paths. We present AegisAsync, a bytecode-based analysis of Python async risks for CPython 3.10-3.14. Its Async Event Transition Graph combines bytecode control flow, resolved calls, async events, and symbolic object relations. Bounded path-local simulation tracks event order and object states and explores cancellation at await suspensions. Six detector plugins evaluate the resulting traces and states to produce candidates with supporting witnesses. On AsyncBench, our benchmark of 180 labeled risks, AegisAsync detects all labeled risks with zero false positives on each of the five versions. Across five path budgets on 2,500 PyPI package versions, source review confirms 457 of 910 AegisAsync candidate records (50.2%), representing 457 unique bugs; across three flake8-async samples, 70 of 2,187 review judgments are confirmed (3.2%). By submission, 166 PoCs reproduce AegisAsync findings, and we have reported 157 bugs to developers, of which 16 have been confirmed and fixed.

cs.CR↗

arXiv2026

DexJoCo-X: Benchmarking Action Representations for Multi-Hand Dexterous Manipulation

As dexterous hands proliferate, collecting data and training policies separately for every morphology becomes increasingly impractical. Scalable cross-embodiment learning therefore requires a unified representation that captures shared manipulation structure while preserving morphology-specific control. Differences in hands, tasks, datasets, and control interfaces prevent existing studies from isolating the effects of representation, pretraining, and architecture. We introduce DexJoCo-X, a benchmark and toolkit for controlled comparison across seven representative dexterous hands, six single-arm and bimanual tasks, and 2,100 balanced demonstrations. DexJoCo-X provides a matched multi-hand, multi-task protocol with common scenes, success criteria, and execution interfaces, redesigned glove-to-hand mappings, and an automated pipeline that expands reviewed demonstrations across randomized scenes. Using $π_{0.5}$, Ego-Pi, and Being-H0.5, we examine whether a shared action interface is sufficient for multi-hand learning. Expanding $π_{0.5}$ to an 80-dimensional bimanual output yields near-zero success. Ego-Pi preserves the pretrained action head through interleaved prediction and supports per-hand multi-task learning, but remains ineffective for seven-hand joint training. By contrast, Being-H0.5 combines cross-embodiment pretraining, a unified action space, and embodiment-aware experts, enabling one policy to control all seven hands. Within this architecture, function-aligned action slots achieve 47.7% mean success, compared with 47.0% for native coordinates and 33.1% for DexLatent. These results show that cross-embodiment representation depends on the entire learning system: action coordinates, pretraining, and architecture must jointly separate shared manipulation structure from embodiment-specific control.

cs.RO↗

arXiv2026

TGuide: Faster Autoregressive Video Diffusion via Historical Trajectory Guidance

Can autoregressive video diffusion be accelerated further when only four denoising steps remain? Existing feature caching approaches struggle in this regime, as large feature changes between denoising timesteps undermine both direct reuse and lightweight predictors. We observe that features across video chunks remain strongly correlated at matching denoising timesteps, offering a complementary source of guidance. Building on this insight, we introduce TGuide, a framework that accelerates autoregressive video diffusion through a lightweight Anchor-Transport-Correct (ATC) predictor guided by historical sampling trajectories. Our ATC combines two complementary references: the current chunk's preceding-step features preserve its spatial structure, while the previous chunk's same-timestep features provide information at the target denoising stage. By integrating these references through spatial retrieval and adaptive correction, ATC improves velocity prediction accuracy despite the large gaps between denoising steps. To mitigate error accumulation across both denoising steps and video chunks, we train the predictor on its own autoregressive rollouts, allowing it to adapt to states shaped by its own earlier predictions. An optional confidence-aware scheduler further adjusts the quality-efficiency trade-off. Experiments on Self Forcing, Causal Forcing, and HY-WorldPlay show that TGuide preserves generation quality while achieving denoising speedups of up to $1.89\times$ over the original 4-step generators. On Self Forcing, TGuide generalizes from 81-frame training videos to 333-frame generation, approximately four times the training length, while exceeding the original model's VBench score.

cs.CV↗

arXiv2026

LightLoc++: Sensor-Robust Representation Learning for Efficient Outdoor LiDAR Localization

Scene coordinate regression (SCR) achieves strong performance in outdoor LiDAR localization, but it usually requires scene-specific training that can take days, limiting practical deployment. Recent works improve training efficiency by decoupling SCR into a scene-agnostic backbone and scene-specific prediction heads, where the backbone is pretrained on source datasets and frozen for new scenes, and only lightweight heads are optimized. However, we find that this paradigm heavily depends on the pretrained backbone. Existing decoupled methods can match conventional SCR methods fully optimized for each new scene when LiDAR configurations are similar to those used during backbone pretraining, but their accuracy drops noticeably on datasets collected with different LiDAR sensors. This suggests that efficient LiDAR localization requires representations that capture stable scene geometry across LiDAR configurations. Motivated by this observation, we propose LightLoc++, a sensor-robust and efficient outdoor LiDAR localization framework. To support sensor-robust representation learning, we introduce SULID, a synchronized urban multi-LiDAR dataset with representative 32-, 64-, and 128-beam rotating LiDARs, extensive cross-sensor overlap, and diverse urban scenes. Using SULID, we pretrain a sensor-robust backbone through cross-sensor consistency learning. LightLoc++ further preserves efficient new-scene learning by incorporating sample classification guidance and redundant sample downsampling, which reduce regression ambiguity and computational redundancy in large-scale outdoor scenes. Extensive experiments on multiple outdoor LiDAR localization benchmarks demonstrate that LightLoc++ achieves state-of-the-art localization performance with the lowest new-scene training cost among compared methods. Code and dataset will be made available at https://github.com/liw95/LightLoc-PlusPlus.

cs.CV↗

arXiv2026

DR-IPC: Disturbance-Resilient Integrated Planning and Control for LiDAR-Based Quadrotor Navigation

LiDAR-based quadrotor navigation in cluttered environments remains challenging under external disturbances, particularly when obstacle-aware motion generation and disturbance-rejection control are handled in separate layers. This article presents disturbance-resilient integrated planning and control (DR-IPC), which combines lightweight path guidance with nonlinear model predictive control (NMPC) to directly generate angular velocity and thrust. An interconnected extended Kalman filter and nonlinear disturbance observer jointly provide filtered state estimates and reconstructed disturbances for NMPC prediction. The resulting formulation unifies nonlinear quadrotor dynamics, actuator constraints, local motion generation, and penalised safe-flight-corridor residuals without requiring a separate trajectory-optimization stage. Gazebo and MARSIM simulations, together with indoor and outdoor experiments, validate DR-IPC under wind, suspended payloads, narrow passages, ball impacts and reactive avoidance of a dynamic obstacle. In multi-goal navigation with disturbances, DR-IPC increases the number of completed missions from 1/10 to 9/10 in Gazebo and reduces the altitude RMSE from 0.34 to 0.01 m in experiments. The complete system operates onboard at 100 Hz. Supplementary videos are available on the project page https://drpp316.github.io/DR-IPC-Page/, and the source code will be released.

cs.RO↗

arXiv2026

Temporal-Aware Fusion for Robust Outdoor LiDAR Localization

LiDAR relocalization aims to estimate the global 6-DoF pose of a sensor in the environment. However, existing regression-based approaches often encounter limitations in dynamic or ambiguous scenarios, as they typically prioritize single-frame inference, leaving the potential of spatio-temporal consistency across scans not fully explored. In this paper, we propose a Temporal-aware Localization framework (TempLoc) designed to enhance the robustness of outdoor localization by effectively modeling sequential consistency. Specifically, a Global Coordinate Estimation module is first introduced to predict point-wise global coordinates and associated uncertainties for each LiDAR scan. A Prior Coordinate Generation module is then presented to estimate inter-frame point correspondences by the attention mechanism. Lastly, an Uncertainty-Guided Coordinate Fusion module is deployed to integrate both predictions of point correspondence in an end-to-end fashion, yielding a more temporally consistent and accurate global 6-DoF pose. Experimental results on the NCLT and Oxford RobotCar benchmarks show that our TempLoc outperforms state-of-the-art methods by a large margin, demonstrating the effectiveness of temporal-aware correspondence modeling in LiDAR relocalization.

cs.CV↗

arXiv2026

CONCURDEP: Event-Guided Analysis of Dependency Invalidation in CPython Concurrency

Removing CPython's Global Interpreter Lock (GIL) exposes native code to concurrency absent from ordinary C types. Mutation or re-entry can revoke a borrowed object, storage pointer, traversal state, or lease between acquisition and use while its owner remains alive, causing native memory errors and runtime-state corruption. Race analyses track conflicting accesses. Python/C lifecycle analyses track individual object states. These reporting units leave implicit owner-subject-storage relations disconnected from later uses under parallel and re-entrant events. We present CONCURDEP, a source-level static analysis of dependency invalidation. Its key insight is to represent the runtime property a native use requires and ask which target-matched event can revoke it within the dependency's live region. CONCURDEP recovers runtime-semantic dependencies, connects them to events through an event-aware native concurrency dependency graph, and applies property-specific state and protection transfers through a shared engine and six mechanism plugins. CONCURDEP correctly classifies all 180 matched semantic-conformance cases and analyzes each of three production CPython releases with five-run medians of 22.13-27.66 seconds and 670-784 MiB peak resident memory. Source auditing confirms 1,094 of 4,273 unique production fingerprints (25.60% confirmation yield). Removing derived relation recovery loses 25-44 represented roots per release; removing cross-entry events loses 73-94. The study identifies 144 distinct bugs across free-threaded and conventional-GIL builds, including 95 previously unreported in public sources. These results show that explicit dependency, event, and property semantics expose consequential runtime failures across API boundaries and execution modes at release scale.

cs.CR↗

arXiv2026

CPyGraph: A Version-Aware Static Analysis Framework for Native CPython Bytecode

Static analysis of Python packages must recover both program structure and object flow across first-class functions, dynamic dispatch, implicit protocol calls, exceptions, closures, and module execution. Native CPython bytecode provides the executable lowering of these behaviors, but its instruction, call, stack, and exception representations change across releases. This creates a need for a version-aware analysis foundation whose graph products share the same bytecode identities and semantics. We present CPyGraph, a C++ framework for package-level analysis of native CPython bytecode. Version- specific adapters expose stack, control, call, lexical, protocol, and exception semantics through a shared interface while preserving code-object identities and native bytecode offsets. An operand-stack-aware Andersen points- to analysis and call graph grow together to a fixed point. Their shared state supports exception-aware CFGs, block-level CDGs, and interprocedural DDGs, with optional function-level flow, context, and bounded path sensitivity. The framework also records unresolved dynamic behavior through typed coverage summaries. We evaluate CPyGraph with PYGBench, 201 package-level programs and 1,733 fixed candidates. On CPython 3.10, the default analysis reaches 91.40% candidate precision and 100% recall; complete sensitivity reaches 94.97% precision with the same recall. Across CPython 3.10-3.14, 1,328 of 1,334 version-invariant queries agree, and CPyGraph matches PyCG on its 112-program call-graph benchmark.

cs.CR↗

arXiv2026

SaltyMeta: a curated benchmark and protein language model-informed web tool for salty peptide prediction

Excess sodium intake remains a major public health challenge, while salty and saltiness-enhancing peptides offer a potential route to preserve sensory saltiness in reduced-sodium foods. Machine-learning studies of salty peptides, however, are constrained by small datasets, heterogeneous evidence standards, uncertain negative labels, and sequence similarity leakage. Here we present SaltyMeta, a curated benchmark and web-accessible screening framework for salty or saltiness-enhancing short peptides. The benchmark contains 580 peptides, including 280 positive peptides and 300 negative peptides, all standardized to 2-15 residue one-letter amino-acid sequences. Quality control found no non-standard residues, exact duplicates, or positive-negative overlaps. A similarity-grouped split retained 456 peptides for training and 124 for held-out testing. We evaluated 548 interpretable peptide descriptors, frozen ESM2 embeddings at 8M, 35M, and 150M parameter scales, and descriptor-embedding fusion models under grouped cross-validation. The traditional ExtraTrees baseline selected by training-set grouped cross-validation achieved ROC-AUC=0.693 in cross-validation and ROC-AUC=0.704, PR-AUC=0.702, F1=0.626, and MCC=0.304 on the held-out test set. The best initial protein-language-model fusion was traditional descriptors plus ESM2-8M embeddings, with grouped CV ROC-AUC=0.696 and test ROC-AUC=0.700. Advanced optimization using PCA95 dimensionality reduction and ExtraTrees feature-importance filtering yielded a practical ESM2-8M PCA95 top-300 model with test ROC-AUC=0.715 and PR-AUC=0.703, although repeated-CV gains remained modest. SaltyMeta is a transparent prioritization tool, not a sensory validation substitute. We provide benchmark, models, scripts, GitHub, and Streamlit for reproducible screening of food-derived peptides.

q-bio.BM↗

arXiv2026

Python Import as an Execution Boundary: An Empirical Study of Bugs, Vulnerabilities, and Analysis Gaps

Python import does more than resolve dependencies: it executes code during module and package initialization. This behavior can trigger failures, load dynamic or native code, access resources, or change security-sensitive state before an application calls a package API. Prior work studies package selection, malicious packages, or package vulnerabilities. We present ImportMine, a study of import-related bugs and security vulnerabilities in Python software. We combine security advisories with PyPI project histories and use source and patch evidence to confirm how import activates cases, why the problem occurs, how developers fix it, and what program information is needed to explain the behavior. We retain 31 import-related advisory vulnerabilities and 38 application-data boundary cases and confirm 1,429 project-history bugs across 1,302 repositories. Among the project-history bugs activated during initialization, 97.6% stop or disrupt normal execution. In contrast, 90.0% of the 20 initialization- activated advisory vulnerabilities are High or Critical. Module-level code and package initialization activate 98.3% of the analyzed history cases. Dynamic loading is much less common, but most of its cases perform security-sensitive actions. We also find that many fixes change when an import becomes active instead of removing the dependency. Finally, we derive ImportVulBench, 228 paired pre-fix and fixed programs covering all 11 bug types.

cs.CR↗

arXiv2026

CellSecInspector: Safeguarding Cellular Networks via Automated Security Analysis on Specifications

The complexity, interdependence, and rapid evolution of 3GPP specifications present fundamental challenges for ensuring the security of modern cellular networks. Manual reviews and existing automated approaches, which often depend on rule-based parsing or small sets of manually crafted security requirements, fail to capture deep semantic dependencies, cross-sentence/clause relationships, and evolving specification behaviors. In this work, we present CellSecInspector, an automated framework for security analysis of 3GPP specifications. CellSecInspector extracts structured state-condition-action (SCA) representations, models mobile network procedures with comprehensive function chains, systematically validates them against 9 foundational security properties under 4 adversarial scenarios, and automatically generates test cases. This end-to-end approach enables the automated discovery of vulnerabilities without relying on manually predefined security requirements or rules. Applying CellSecInspector to the well-studied 5G and 4G NAS and RRC specifications and selected sections of TS 23.501 and TS 24.229, it discovers 43 vulnerabilities, 7 of which are previously unreported. Our findings show that CellSecInspector is a scalable, adaptive, and effective solution to assess 3GPP specifications for safeguarding operational and next-generation cellular networks.

cs.CR↗

arXiv2026

Learning to Use Imagination: Progress-Conditioned Future Utilization for World Action Models

World Action Models (WAMs) extend Vision-Language-Action (VLA) models by incorporating future visual dynamics into action generation. However, existing WAMs often utilize imagined futures with limited adaptation to evolving execution progress, potentially introducing distracting or unreliable predictive cues. This limitation arises from two empirically identified forms of non-uniformity in future utility: (i) at the inter-progress level, the utility of imagined futures varies across execution stages as control demands change; and (ii) at the intra-progress level, individual future latents exhibit heterogeneous relevance within the same progress state. To address these limitations, we propose ProWAM, a Progress-Conditioned World Action Model that introduces execution progress as an explicit intermediate representation for adaptive imagination utilization. ProWAM comprises two tightly coupled components: (1) To obtain a reliable representation of execution progress, we propose the Self-Supervised Dual-Temporal Progress Encoder (SS-DTPE). SS-DTPE couples short-term action-observation interaction modeling with long-term recurrent progress aggregation to capture recent execution feedback and accumulated task history. (2) Conditioned on the progress representation from SS-DTPE, we propose the Hierarchical Progress-Conditioned Imagination Modulation (HPIM) to adapt imagination utilization to execution progress. HPIM operates at two complementary levels: an inter-progress global modulation mechanism adapts future utilization across execution stages, while an intra-progress relevance mechanism differentiates individual future latents within each progress state. Extensive experiments demonstrate consistent gains over strong VLA and WAM baselines.

cs.CV↗

arXiv2026

Physically Plausible Video Generation via Visual-Semantic Chain-of-Events Conditioning

Physically Plausible Video Generation (PPVG) seeks to synthesize videos consistent with physical principles, yet remains challenging due to underspecified natural language conditioning. Advanced chain-of-thought (CoT) frameworks augment prompts with physical knowledge. However, such prompts describe physical phenomena holistically, overlooking intermediate states and transition dynamics. In this paper, we reformulate PPVG as event-centric generation by representing physical evolution as a chain of causally connected and physically constrained events. Our framework comprises three key modules: (1) Physics-driven Event Chain Reasoning. This module decomposes physical phenomena into causally connected events represented by evolving scene graphs. Formula-derived physical quantities are bound to relevant objects and interactions, characterizing the direction and magnitude of each event transition. (2) Transition-aware Routed Keyframe Conditioning. This module routes each event to a specialized keyframe synthesis operator for appearance variation or object transformation. Consecutive keyframes are injected as residual guidance during denoising, enabling smooth visual transitions between event-boundary states. (3) Physics-injected Contrastive Semantic Guidance. This module constructs physics-informed positive and counterfactual negative prompts for classifier-free guidance, steering generation toward plausible dynamics and away from physics-violating counterparts. Experiments on PhyGenBench, VideoPhy, PhyWorldBench, and Physics-IQ demonstrate that our framework generates videos with superior physical plausibility across diverse domains.

cs.CV↗

arXiv2026

Physics-Assisted Deep Learning Denoising for Stabilized IMPULSED dMRI Microenvironment Parameter Fitting

Diffusion-weighted MRI (dMRI) is a powerful tool for quantifying cellular microenvironment parameters. This study proposes a physics-assisted deep learning (DL)-based denoising framework designed to enhance dMRI signal quality and improve the robustness of subsequent biophysical model fitting. A dataset of paired noise-free and Rician-noise-corrupted dMRI signals was generated using the IMPULSED-dMRI signal model. Three denoising architectures were evaluated: Convolutional Neural Networks (CNN), Multilayer Perceptron (MLP), and Long Short-Term Memory (LSTM) networks. Denoised signals were then fitted to estimate cell diameter $d$, intracellular volume fraction $V_{\mathrm{in}}$, and extracellular apparent diffusion coefficient $D_\mathrm{ex}$. DL-based processing substantially improved dMRI signal denoising. The MLP and LSTM achieved similar performance, with the LSTM slightly better overall, and both outperformed the CNN. In the subsequent model fitting step, the LSTM produced modest reductions in parameter MAE. The dominant benefit was fitting stabilization, with the overall fitting failure rate reduced from 57.6\% to 17.7\%. The proposed framework improves dMRI signal quality and stabilizes subsequent IMPULSED-based microenvironmental parameter fitting.

physics.med-ph↗