arXiv subjects

Jason Wu

Publications and source records attributed to Jason Wu.

At least 19 recordsLinked to original sources

arXiv2026

Revision-Based Learning from Interactive Experience and Feedback

Code generation models are increasingly used to produce interactive artifacts, such as websites and user interfaces. Many properties that determine whether these interfaces are useful only become apparent through realistic interaction, making them difficult to incorporate into model training. In this paper, we introduce RELIEF (REvision-based Learning from Interactive Experience and Feedback), a method for improving code generation models using feedback on their generated interfaces. RELIEF uses a coding agent to revise these interfaces, then incorporates the improvements into their original generation traces to produce supervision for fine-tuning. To evaluate RELIEF, we collected a dataset of 200 synthetically generated websites paired with 355 interaction recordings and accompanying feedback from 24 users. We used critiques derived from code inspection, automated interaction, explicit human feedback, and human interaction recordings to fine-tune variants of two code generation models. Our results show that feedback effectiveness varies with base model capabilities, with critiques inferred from human interaction recordings leading to the largest usability improvements overall under our tested models.

cs.AI↗

arXiv2026

EGI-BENCH: An Enterprise General Intelligence Benchmark Suite

Enterprise adoption of LLMs in agentic workflows is gated by evaluations that span a wide arc of abilities: CRM operations, heterogeneous-document retrieval, personal data understanding, multi-turn planning, tool use, function calling, long-horizon research, and long-context software engineering. These evaluations live today in separate repositories with divergent runners, dependencies, and scoring conventions, which makes cross-benchmark comparison difficult and obscures what a single aggregate number actually represents. We release EGI-BENCH, a unified LLM agent benchmark suite that consolidates eight enterprise facing agent benchmarks (CRMARENA, HERB, PERSONABENCH, USERBENCH, MCP-UNIVERSE, MCPEVAL, BENCHFORCE, LOCOBENCH) behind a single command-line interface with two runtime-budgeted execution modes. On top of the suite we define the Enterprise General Intelligence (EGI) Score, a six-dimensional framework (M1-M6) whose buckets are populated by fine-grained sub-metrics routed from each benchmark rather than by benchmark-level averages. We describe the suite, the sub-metric routing, and the quick/deep budgets, and report results for fourteen frontier and cost-efficient LLMs spanning the GPT, Claude, and Gemini families under quick mode. Top-line EGI ranges from 39.6 (GPT-4o mini) to 65.6 (gpt-5.5), and no single model dominates every capability axis: gpt-5.5 leads task outcome, planning, and retrieval grounding; Claude Sonnet 4.5 leads tool actioning and safety/abstention; and Claude Haiku 4.5 leads engineering quality. Code is available at https://github.com/SalesforceAIResearch/egi-bench.

cs.AI↗

arXiv2026

ARISE: Adaptive Agentic Reasoning with Image-grounded Self-Evaluation for Interpretable IBD Assessment

Inflammatory bowel disease (IBD) requires frequent imaging-based assessment, yet interpretation of modalities such as wireless capsule endoscopy (WCE) and intestinal ultrasound remains heavily dependent on specialist expertise. Vision-Language Models (VLMs) demonstrate significant potential in multimodal medical image analysis, but their clinical adoption is hindered by their insufficient domain-specific reasoning, susceptibility to hallucination, scarcity of high quality training data in fine-grained diagnostics and limited interpretability. We introduce ARISE (Adaptive Agentic Reasoning with Image-grounded Self-Evaluation), an autonomous planning framework that models few-shot medical image understanding as a sequential agentic workflow. ARISE structures agent execution into a transparent 5-stage workflow: hypothesis generation, image-grounded evidence summarization, evidence-conditioned refinement, symbolic verification, and final diagnosis. We apply ARISE to IBD assessment across two independent patient cohorts: wireless capsule endoscopy (WCE) images for Crohn's disease and B-mode ultrasound data for ulcerative colitis. ARISE consistently improves diagnostic performance over baseline VLMs while exposing where reasoning succeeds or fails, providing a more interpretable basis for clinical decision support and realistic deployment.

cs.CV↗

arXiv2026

GUIDE: Designer-in-the-loop Authoring of Conformant Generative User Interfaces

Generative User Interfaces (GenUIs) enable applications to generate interfaces on demand from user needs and context. Like conventional UIs, they must still reflect designers' intent and conform to requirements such as brand identity. Unlike conventional UIs, designers cannot directly specify or see every interface a GenUI may produce, making design intent harder to enforce. We introduce GUIDE (GenUI Development Environment), a system that lets designers continuously inspect and refine GenUI behavior as they create and edit interfaces. GUIDE uses designers' modifications and interactions to adapt GenUIs through prompt optimization and a novel adaptive conformance scoring model. We validate GUIDE's scoring model and system. The scoring model matched or outperformed proprietary LLM baselines on held-out comparisons of real and synthetic application screens. In a study with 12 UI/UX practitioners, participants found GUIDE effective and usable and significantly preferred aligned outputs over a strong baseline using exemplars and a model-generated design.md.

cs.HC↗

arXiv2026

Efficient Personalization of Generative User Interfaces

Generative user interfaces (GenUIs) create new opportunities to adapt interfaces to individual users on demand. Yet personalization is difficult because it is not possible to provide settings for screens that have not yet been generated, making it necessary to learn preferences from users' feedback on generated interfaces. Such feedback is sparse, subjective, and often difficult to articulate. We study this problem through a new dataset in which 20 participants each judge the same 600 pairs of GenUIs. Their judgments differ substantially (Krippendorff's alpha = 0.25), and written rationales show that even when participants consider similar UI properties, such as color tone or hierarchy, they often make different choices and prioritize them differently. We develop a sample-efficient personalization method that uses a few pairwise judgments to weight prior users' preferences rather than relying on a fixed rubric of UI attributes. Our method outperforms a pretrained UI evaluator and a larger multimodal model offline. In an online study with 12 new users, participants preferred interfaces personalized by our method over all baselines, including shared guidelines and users' own written preferences.

cs.LG↗

arXiv2026

The EDD Radio Astronomy Backend Framework

Modern digital radio astronomy receivers produce increasingly wide-bandwidth, high bit-rate data streams that necessitate the development of flexible, scalable, and maintainable backend processing and recording systems. Historically, such backend instrumentation has been tightly coupled to telescope observing modes, limiting reuse between observatories and science cases. We present the Effelsberg Direct Digitisation (EDD) backend framework, a software-defined architecture for constructing real-time radio astronomy backends on commodity off-the-shelf computing infrastructure. We describe its design, implementation, supported observing modes, and operational deployments. EDD separates a common core framework from plugin-provided observing capabilities. The core provides orchestration, telescope interfaces, pipeline lifecycle management, monitoring, and deployment tooling, while plugins implement processing pipelines for specific observing modes. The framework is designed to support both single-dish and interferometric instruments through site-specific configuration and plugin selection. EDD currently supports spectroscopy and spectropolarimetry, pulsar timing and searching, baseband recording, very long baseline interferometry, correlation, and beamforming. Operational deployments include the Effelsberg 100-m telescope, the SKA-MPI prototype dish, the Thai National Radio Telescope, and the ARGOS interferometric prototype array. By separating common services, observing-mode plugins, and site-specific configuration, it allows backend capabilities to be deployed across heterogeneous telescope environments and provides a community resource for broadband radio astronomy instrumentation.

astro-ph.IM↗

arXiv2026

Range-Aware Bayesian Optimization for Discovering Diverse Designs within Target Property Windows

In many materials and product design problems, desirable candidates exhibit properties that fall within an acceptable range rather than achieve a single optimum. Recovering multiple, distinct solutions that satisfy such specifications is also practically valuable, as some candidates may be preferred for reasons of cost, processability, or robustness that are difficult to encode directly in an objective function. Here, we develop a range-aware Bayesian optimization (BO) framework in which the acquisition function directly scores the posterior probability that a candidate satisfies a target range. The framework naturally extends to parallel pursuit of multiple distinct specifications over a shared candidate space. Across benchmark tasks, range-aware acquisition consistently recovers larger and more diverse sets of valid designs than standard BO baselines and recent goal-seeking methods. Its utility is further demonstrated in two practically motivated design case studies involving optimizing reaction conditions for polymer synthesis and sequence-defined oligomer discovery for prescribed optical absorption bands, supported by quantum chemical calculations. These results suggest that range-aware BO can provide a practical and sample-efficient foundation for specification-driven design, particularly when design flexibility and solution diversity are important considerations.

cs.LG↗

arXiv2026

TS-Skill: A Benchmark for Evaluating Analytical Skills in Time-Series Question Answering

Large language models (LLMs) and time-series language models (TSLMs) are increasingly applied to time-series question answering (TSQA). Unlike text-only QA, TSQA requires models to ground answers in temporal signals whose patterns may occur at different scales, specific time locations, or across separated intervals. However, existing benchmarks are typically organized by task types or high-level reasoning categories, making it difficult to diagnose the underlying signal-level capabilities driving model performance. We introduce TS-Skill, a controlled benchmark for evaluating three composable analytical skills in TSQA: temporal scale selection (SK1), temporal localization (SK2), and cross-interval integration (SK3). TS-Skill provides timestamp-aware questions, broad domain coverage, and human-validated QA quality. To construct the benchmark at scale, we develop SKEvol, a skill-guided agentic framework that combines domain-aware time-series seed generation, skill-controlled question generation, metadata- and code-assisted answer construction, multi-phase signal-grounded verification, and human-in-the-loop curation. Experiments on ten state-of-the-art LLMs and TSLMs reveal substantial and uneven capability gaps across SK1-SK3. In particular, SK3 remains consistently challenging for non-agent models, whereas tool-augmented agents show a selective advantage on standalone SK3. These findings demonstrate that skill-level evaluation can uncover temporal reasoning failures that are obscured by aggregate TSQA scores.

cs.CL↗

arXiv2026

FlowEval: Reference-based Evaluation of Generated User Interfaces

While large language models (LLMs) and coding agents are often applied to user interface (UI) development, developers find it difficult to reliably assess their proficiency in visual and interaction design. Existing evaluations either rely on human experts, who can accurately assess usability by testing critical flows but are slow and costly, or on automated judges, which are scalable but less accurate and opaque. We present FlowEval, a reference-based framework that measures whether a generated UI supports realistic interaction flows by comparing navigation traces from real websites to traces from generated analogs using reference-based similarity metrics (e.g., dynamic time warping). In a small-scale study with expert UI evaluators, we show that reference-based metrics strongly correlate with human judgments, suggesting that they can provide scalable yet trustworthy evaluation for UI generation systems.

cs.MA↗

arXiv2026

SWAN: World-Aware Adaptive Multimodal Networks for Runtime Variations

Multimodal deep neural networks deployed in realistic environments must contend with runtime variations: changes in modality quality, overall input complexity, and available platform resources. Current networks struggle with such fluctuations -- adaptive networks cannot adhere to a strict compute budget, controller-based networks neglect to consider input complexity, and statically provisioned networks fail at all the above. Consequently, they do not extract maximum utility from the expended computational resources. We present SWAN (Sample and World-Aware Multimodal Network), the first adaptive multimodal network that accomplishes all three goals. SWAN employs a quality-aware controller to assign resources among modalities according to a variable user-specified maximum budget. Within this budget, an adaptive gating module further optimizes efficiency by scaling layer utilization according to sample complexity. For further gains, SWAN also employs a token dropping module that masks semantically irrelevant multimodal features before performing detections. We evaluate SWAN in the domain of autonomous driving with complex multi-object 3D detection, reducing FLOPs by up to 49% with minimal degradation.

cs.LG↗

arXiv2026

Training Computer Use Agents to Assess the Usability of Graphical User Interfaces

Usability testing with experts and potential users can assess the effectiveness, efficiency, and user satisfaction of graphical user interfaces (GUIs) but doing so remains a costly and time-intensive process. Prior work has used computer use agents (CUAs) and other generative agents that can simulate user interactions and preference, but we show that agents still struggle to provide accurate usability assessments. In this work, we present a novel machine learning method that operationalizes a computational definition of usability to train CUAs to assess GUI usability by i) prioritizing important interaction flows, ii) executing them through human-like interactions, and iii) predicting a learned numerical usability score. We train a computer use agent, uxCUA, with our algorithm on a large-scale dataset of fully interactive user interfaces (UIs) paired with usability labels and human preferences. We show that uxCUA outperforms larger models in accurate usability assessments and produces realistic critiques of both synthetic and real UIs. More broadly, our work aims to build a principled, data-driven foundation for automated usability assessment in HCI.

cs.CL↗

arXiv2026

Decoupling Vision and Language: Codebook Anchored Visual Adaptation

Large Vision-Language Models (LVLMs) use their vision encoders to translate images into representations for downstream reasoning, but the encoders often underperform in domain-specific visual tasks such as medical image diagnosis or fine-grained classification, where representation errors can cascade through the language model, leading to incorrect responses. Existing adaptation methods modify the continuous feature interface between encoder and language model through projector tuning or other parameter-efficient updates, which still couples the two components and requires re-alignment whenever the encoder changes. We introduce CRAFT (Codebook RegulAted Fine-Tuning), a lightweight method that fine-tunes the encoder using a discrete codebook that anchors visual representations to a stable token space, achieving domain adaptation without modifying other parts of the model. This decoupled design allows the adapted encoder to seamlessly boost the performance of LVLMs with different language architectures, as long as they share the same codebook. Empirically, CRAFT achieves an average gain of 13.51% across 10 domain-specific benchmarks such as VQARAD and PlantVillage, while preserving the LLM's linguistic capabilities and outperforming peer methods that operate on continuous tokens.

cs.CV↗

arXiv2026

Improving User Interface Generation Models from Designer Feedback

Despite being trained on vast amounts of data, most LLMs are unable to reliably generate well-designed UIs. Designer feedback is essential to improving performance on UI generation; however, we find that existing RLHF methods based on ratings or rankings are not well-aligned with with designers' workflows and ignore the rich rationale used to critique and improve UI designs. In this paper, we investigate several approaches for designers to give feedback to UI generation models, using familiar interactions such as commenting, sketching and direct manipulation. We first perform an evaluation with 21 designers where they gave feedback using these interactions, which resulted in 1500 design annotations. We then use this data to finetune a series of LLMs to generate higher quality UIs. Finally, we evaluate these models with human judges, and we find that our designer-aligned approaches outperform models trained with traditional ranking feedback and all tested baselines, including GPT-5.

cs.HC↗

arXiv2026

Detection of Cyclotron Absorption in the Radio Emission of GPM 1839-10

GPM 1839-10 is an intriguing long-period radio transient (LPT), distinguished by its activity spanning at least three decades and its highly unusual emission characteristics. These features include orthogonal polarization mode (OPM) switches, down-drifting sub-structures, and distinct linear-to-circular polarization conversion behaviors. In this work, we present follow-up observations utilizing the FAST telescope at L-band, yielding a total of seven detected radio pulses. We find a consistent association between OPM switches and a decrease in polarized intensity. This feature strongly supports the hypothesis that the OPM switches are generated by the incoherent summation of OPMs. Our measured Rotation Measures (RMs) are consistent with previous observations, indicating that the magneto-ionic environment is stable. If the source is in a binary system, such stability suggests it may host a weakly magnetized companion. Crucially, we firstly observe clear evidence of a cyclotron absorption feature in one radio pulse, a signature rarely observed in radio sources. This feature allows us to infer that the magnetic field strength at the absorption site has a lower limit of tens of Gauss, which is necessary for the phenomenon to occur. This characteristic can be explained in a scenario where GPM 1839-10 possesses a weakly magnetized companion star.

astro-ph.HE↗

arXiv2025

STAR: Smartphone-analogous Typing in Augmented Reality

While text entry is an essential and frequent task in Augmented Reality (AR) applications, devising an efficient and easy-to-use text entry method for AR remains an open challenge. This research presents STAR, a smartphone-analogous AR text entry technique that leverages a user's familiarity with smartphone two-thumb typing. With STAR, a user performs thumb typing on a virtual QWERTY keyboard that is overlain on the skin of their hands. During an evaluation study of STAR, participants achieved a mean typing speed of 21.9 WPM (i.e., 56% of their smartphone typing speed), and a mean error rate of 0.3% after 30 minutes of practice. We further analyze the major factors implicated in the performance gap between STAR and smartphone typing, and discuss ways this gap could be narrowed.

cs.HC↗

arXiv2025

DesignPref: Capturing Personal Preferences in Visual Design Generation

Generative models, such as large language models and text-to-image diffusion models, are increasingly used to create visual designs like user interfaces (UIs) and presentation slides. Finetuning and benchmarking these generative models have often relied on datasets of human-annotated design preferences. Yet, due to the subjective and highly personalized nature of visual design, preference varies widely among individuals. In this paper, we study this problem by introducing DesignPref, a dataset of 12k pairwise comparisons of UI design generation annotated by 20 professional designers with multi-level preference ratings. We found that among trained designers, substantial levels of disagreement exist (Krippendorff's alpha = 0.25 for binary preferences). Natural language rationales provided by these designers indicate that disagreements stem from differing perceptions of various design aspect importance and individual preferences. With DesignPref, we demonstrate that traditional majority-voting methods for training aggregated judge models often do not accurately reflect individual preferences. To address this challenge, we investigate multiple personalization strategies, particularly fine-tuning or incorporating designer-specific annotations into RAG pipelines. Our results show that personalized models consistently outperform aggregated baseline models in predicting individual designers' preferences, even when using 20 times fewer examples. Our work provides the first dataset to study personalized visual design evaluation and support future research into modeling individual design taste.

cs.CV↗

arXiv2025

An RFSoC-based F-engine for ARGOS

Radio interferometers provide the means to perform the wide-field-of-view (FoV), high-sensitivity observations required for modern radio surveys. As computing power per cost has decreased, there has been a move towards larger arrays of smaller dishes, such as DSA-2000, the upcoming HIRAX, CHORD and SKA radio telescopes. Such arrays can have simpler receiver designs with room-temperature low-noise amplifiers and direct sampling to achieve, greatly reducing the cost per antenna. The ARGOS project is currently developing an array of five 6-meter antennas that will be used to demonstrate the technology required for a next generation "small-D, big-N" radio interferometer in Europe. In this work, our objective was to implement a first-stage digital signal processing system for the ARGOS demonstrator array, providing digitization, channelization, delay correction and frequency-dependent complex gain correction. The system is intended to produce delay and phase corrected dual-polarization channelized voltages in the frequency range 1-3 GHz with a nominal channel bandwidth of 1 MHz. We use an RFSoC 4x2 evaluation board with four analog-to-digital converters (ADCs) that can simultaneously sample two 1 GHz, dual-polarization bands. We use Xilinx Vitis HLS C++ to develop the required firmware as a set of customizable modules suitable for rapid prototyping. We performed hardware verification of the channel response of the critically sampled PFB and of the delay correction, showing both to be consistent with theoretical expectations. Furthermore, the board was installed at the Effelsberg 100-meter radio telescope where we performed commensal pulsar observations with the Effelsberg Direct Digitization backend, showing comparable performance. This work demonstrates the utility of high-level synthesis (HLS) languages in the development of high performance radio astronomy processing backends.

astro-ph.IM↗

arXiv2025

ADMN: A Layer-Wise Adaptive Multimodal Network for Dynamic Input Noise and Compute Resources

Multimodal deep learning systems are deployed in dynamic scenarios due to the robustness afforded by multiple sensing modalities. Nevertheless, they struggle with varying compute resource availability (due to multi-tenancy, device heterogeneity, etc.) and fluctuating quality of inputs (from sensor feed corruption, environmental noise, etc.). Statically provisioned multimodal systems cannot adapt when compute resources change over time, while existing dynamic networks struggle with strict compute budgets. Additionally, both systems often neglect the impact of variations in modality quality. Consequently, modalities suffering substantial corruption may needlessly consume resources better allocated towards other modalities. We propose ADMN, a layer-wise Adaptive Depth Multimodal Network capable of tackling both challenges: it adjusts the total number of active layers across all modalities to meet strict compute resource constraints and continually reallocates layers across input modalities according to their modality quality. Our evaluations showcase ADMN can match the accuracy of state-of-the-art networks while reducing up to 75% of their floating-point operations.

cs.LG↗