arXiv subjects

Daniel Wang

Publications and source records attributed to Daniel Wang.

At least 19 recordsLinked to original sources

arXiv2026

All parallel chip-firing games with $2|E|-|V|<|\sigma|<2|E|$ have period $2$

In 2010, Levine found that the activity functions of parallel chip-firing games on the complete graph $K_n$ converge to a devil's staircase pattern: adding chips causes activity to progress through open intervals in which it is locally constant. In 2022, Bu, Choi, and Xu improved on an earlier bound by Kominers and Kominers to show that there exists a strict lower bound below which all games have activity $0$, and a strict upper bound above which all games have activity $1$. They thereby generalized the bottom and topmost rungs of the devil's staircase to all graphs. In 2024, Ji, Li, and Wang conjectured that a similarly general bound exists for games with activity $\tfrac12$. We use GPT-5.6-Sol to prove this conjecture, unifying existing results for trees, cycles, complete graphs, and complete bipartite graphs. This generalizes the middle rung of the devil's staircase.

math.CO↗

arXiv2026

AllocBench: Measuring Online Tool Allocation Capability in LLM Agents

Creating a reusable tool is an investment: an agent pays a fixed cost now in exchange for the potential of future reuse. Therefore, a user should prefer an agent that creates a small number of highly reusable tools, rather than many one-offs. We introduce a paired benchmark that tests whether LLM agents exhibit conscious allocation behavior under a fixed budget in two contexts: an abstract text-based formulation and a code-construction task. We find that every frontier model we test---Claude Haiku, Claude Opus, GPT-5.4-mini, and GPT-5.6 Sol---acts near-optimally in the abstract framing but fails to transfer this ability to script-writing. Through further experiments, we identify the particular failure modes for each model. Notably, the first three models fail even when the scripts are not evaluated, while GPT-5.6 Sol stays selective under that weaker manipulation and collapses only at full construction. Furthermore, an open-source Qwen model policy-trained for abstract allocation generalizes this ability across held-out lexical variations, but sees no improvement at script allocation. Together, these results establish online tool allocation as a significant capability boundary, even for modern frontier models.

cs.LG↗

arXiv2026

Lomekwi: Resource-Bounded Tool Discovery in LLM Agents

Existing tool-use benchmarks report a single success rate for complex, multistep tasks. Inspired by ideas from cognitive science, we distinguish tool use from tool discovery and decompose the latter into curiosity (the model's ability to discover the parts needed to build the tool), recognition (the model's ability to discover the process of creating the tool), and efficiency (the model's use of the tool after creation). We show that this framework can be applied to existing discovery tasks, such as Voyager. In addition, we provide evidence that recognition inversely scales with model size, and we introduce and analyze a class of combinatorial games that demonstrates this. We further observe inverse scaling in a separate environment designed to emulate real-world tasks.

cs.AI↗

arXiv2026

Biased-Noise Quantum Reed-Solomon Codes and a Tornado Concatenation for Cat Qubits

Dissipative cat qubits exponentially suppress one Pauli error channel with the mean photon number, leaving the conjugate bit-flip error as the dominant failure mode. This strong noise bias makes the full machinery of general quantum error correction unnecessary: a code need only protect against a single error type, and any classical linear code can be promoted to a Clifford stabilizer code that does exactly this. We use this observation to build a bit-flip-only quantum Reed-Solomon (RS) code. Starting from the maximum-distance-separable RS code [7, 3, 5] over GF($2^3$), we binary-expand it to the linear code [21, 9, 6] over GF(2) and realize it as a [[21, 9, $d_X = 6, d_Z = 1$]] bit-flip code whose stabilizers are products of $Z$ operators. Because no phase-flip correction is attempted, the construction discards the redundancy that standard quantum RS codes spend on correcting $Z$ errors -- which a strongly biased cat qubit renders unnecessary -- and yields a shallow Clifford circuit that samples directly in Stim. Errors are decoded by an optimal bounded-distance syndrome-lookup table. We then introduce a Tornado architecture: a two-layer concatenation that wraps every position of the outer RS code in an inner distance-three repetition code, yielding a [[63, 9, 18]] code decoded by a two-stage inner majority vote and outer lookup decoder. Monte-Carlo simulation shows that at a physical bit-flip rate $p = 0.1$ the Tornado code reaches a logical error rate $p_L \approx 5.3 \times 10^{-3}$, below both parent codes, and that its logical error rate scales as $p_L \propto p^6$ at low $p$, in contrast to $p^2$ for the repetition code and $p^3$ for the standalone RS code. We give the exact construction, the error and circuit model, an asymptotic scaling analysis, and an honest account of the overhead cost and single-shot assumptions.

quant-ph↗

arXiv2026

SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work?

AI agents are increasingly expected to complete long-horizon workflows that require sustained progress over hours, millions of tokens, and complex environments. Yet current agent benchmarks largely evaluate short-form tasks, such as single pull requests, small tickets, or 5-10 minute exercises, limiting our ability to measure agents' capabilities in planning, long-context understanding, and memory use. We introduce SWE-Marathon, a benchmark of 20 long-horizon tasks spanning software engineering and adjacent technical domains. Each task consists of a unique executable environment, a human-written reference solution, and a multi-layer verification suite. Logged agent attempts average 27.2M total tokens, making SWE-Marathon substantially longer-horizon than existing SWE and command-line agent benchmarks. Current frontier coding agents solve fewer than 30% of tasks. Failures often arise from poor self-verification, self-reported infeasibility, and premature termination. We also observe reward-hacking behavior in 13.8% of rollouts, where agents attempt to exploit the environment or verifier to bypass the intended workflow. SWE-Marathon includes adversarial review of test suites and execution environments, as well as multi-layer checks designed to prevent shortcut solutions. We release SWE-Marathon, evaluation code, and agent trajectories at https://swe-marathon.org/.

cs.SE↗

arXiv2026

Beyond ReinMax: Low-Variance Gradient Estimators for Discrete Latent Variables

Machine learning models involving discrete latent variables require gradient estimators to facilitate backpropagation in a computationally efficient manner. The most recent addition to the Straight-Through family of estimators, ReinMax, can be viewed from a numerical ODE perspective as incorporating an approximation via Heun's method to reduce bias, but at the cost of high variance. In this work, we introduce the ReinMax-Rao and ReinMax-CV estimators which incorporate Rao-Blackwellisation and control variate techniques into ReinMax to reduce its variance. Our estimators demonstrate superior performance on training variational autoencoders with discrete latent spaces. Furthermore, we investigate the possibility of leveraging alternative numerical methods for constructing more accurate gradient approximations and present an alternative view of ReinMax from a simpler numerical integration perspective.

stat.ML↗

arXiv2026

A Microservice-Based Platform for Sustainable and Intelligent SLO Fulfilment and Service Management

The Microservices Architecture (MSA) design pattern has become a staple for modern applications, allowing functionalities to be divided across fine-grained microservices, fostering reusability, distribution, and interoperability. As MSA-based applications are deployed to the Computing Continuum (CC), meeting their Service Level Objectives (SLOs) becomes a challenge. Trading off performance and sustainability SLOs is especially challenging. This challenge can be addressed with intelligent decision systems, able to reconfigure the services during runtime to meet the SLOs. However, developing these agents while adhering to the MSA pattern is complex, especially because CC providers, who have key know-how and information to fulfill these SLOs, must comply with the privacy requirements of application developers. This work presents the Carbon-Aware SLO and Control plAtform (CASCA), an open-source MSA-based platform that allows CC providers to reconfigure services and fulfill their SLOs while maintaining the privacy of developers. CASCA is architected to be highly reusable, distributable, and easy to use, extend, and modify. CASCA has been evaluated in a real CC testbed for a media streaming service, where decision systems implemented in Bash, Rust, and Python successfully reconfigured the service, unaffected by upholding privacy.

cs.SE↗

arXiv2025

Towards Scalable Backpropagation-Free Gradient Estimation

While backpropagation--reverse-mode automatic differentiation--has been extraordinarily successful in deep learning, it requires two passes (forward and backward) through the neural network and the storage of intermediate activations. Existing gradient estimation methods that instead use forward-mode automatic differentiation struggle to scale beyond small networks due to the high variance of the estimates. Efforts to mitigate this have so far introduced significant bias to the estimates, reducing their utility. We introduce a gradient estimation approach that reduces both bias and variance by manipulating upstream Jacobian matrices when computing guess directions. It shows promising results and has the potential to scale to larger networks, indeed performing better as the network width is increased. Our understanding of this method is facilitated by analyses of bias and variance, and their connection to the low-dimensional structure of neural network gradients.

cs.LG↗

arXiv2025

Fine Tuning Methods for Low-resource Languages

The rise of Large Language Models has not been inclusive of all cultures. The models are mostly trained on English texts and culture which makes them underperform in other languages and cultural contexts. By developing a generalizable method for preparing culturally relevant datasets and post-training the Gemma 2 model, this project aimed to increase the performance of Gemma 2 for an underrepresented language and showcase how others can do the same to unlock the power of Generative AI in their country and preserve their cultural heritage.

cs.CL↗

arXiv2025

Measuring Stereotype and Deviation Biases in Large Language Models

Large language models (LLMs) are widely applied across diverse domains, raising concerns about their limitations and potential risks. In this study, we investigate two types of bias that LLMs may display: stereotype bias and deviation bias. Stereotype bias refers to when LLMs consistently associate specific traits with a particular demographic group. Deviation bias reflects the disparity between the demographic distributions extracted from LLM-generated content and real-world demographic distributions. By asking four advanced LLMs to generate profiles of individuals, we examine the associations between each demographic group and attributes such as political affiliation, religion, and sexual orientation. Our experimental results show that all examined LLMs exhibit both significant stereotype bias and deviation bias towards multiple groups. Our findings uncover the biases that occur when LLMs infer user attributes and shed light on the potential harms of LLM-generated outputs.

cs.CL↗

arXiv2025

VRSight: An AI-Driven Scene Description System to Improve Virtual Reality Accessibility for Blind People

Virtual Reality (VR) is inaccessible to blind people. While research has investigated many techniques to enhance VR accessibility, they require additional developer effort to integrate. As such, most mainstream VR apps remain inaccessible as the industry de-prioritizes accessibility. We present VRSight, an end-to-end system that recognizes VR scenes post hoc through a set of AI models (e.g., object detection, depth estimation, LLM-based atmosphere interpretation) and generates tone-based, spatial audio feedback, empowering blind users to interact in VR without developer intervention. To enable virtual element detection, we further contribute DISCOVR, a VR dataset consisting of 30 virtual object classes from 17 social VR apps, substituting real-world datasets that remain not applicable to VR contexts. Nine participants used VRSight to explore an off-the-shelf VR app (Rec Room), demonstrating its effectiveness in facilitating social tasks like avatar awareness and available seat identification.

cs.HC↗

arXiv2025

Hiding behind a curtain of dust: Gas and dust properties of an ultra-luminous strongly-lensed z = 3.75 galaxy behind the Milky Way disk

We present a detailed analysis of J154506, a strongly lensed submillimeter galaxy behind the Lupus-I molecular cloud, and characterisation of its physical properties using a combination of new and archival data, including VLT/MUSE and FORS2 optical data. We identify two high-significance (SNR>5) emission lines at 97.0 and 145.5 GHz, corresponding to CO(4-3) and CO(6-5), respectively, in the spectral scans from the Atacama Compact Array and the Large Millimetre Telescope and the [CII] 158~$\mu$m fine-structure line at 400~GHz using the Atacama Pathfinder Experiment. These detections yield a spectroscopic redshift of $z_{\rm{spec}}=3.7515\pm0.0005$. We also report the detection of [CI], HCN(4-3), and two H$_2\rm{O}^+$ transitions, further confirming the redshift and providing insights into J154506's physical properties. By modeling sub-arcsecond resolution (0.75) ALMA Band 6 and 7 continuum data in the uv-plane, we derive an average magnification factor of $6.0\pm0.4$ and our analysis reveals a relatively cold dust (38K) in a starburst ($\sim900~\rm{M}_{\odot}yr^{-1}$) galaxy with a high intrinsic dust mass ($\sim2.5\times10^{9}~\rm{M}_{\odot}$) and infrared (IR) luminosity ($\sim6\times10^{12}~\rm{L}_{\odot}$). The non-local thermodynamic equilibrium radiative transfer modelling of the joint dust SED and CO line excitation suggests the dust continuum emission is primarily associated with relatively diffuse regions with molecular gas densities of $10^2-10^4\rm{cm}^{-3}$, rather than compact, high-pressure environments typical of extreme starbursts or AGNs. This is supported by the close-to-unity ratio between the dust and gas kinetic temperatures, which argues against highly energetic heating mechanisms. The CO excitation ladder peaks close to CO(5-4) and is dominated by slightly denser molecular gas.

astro-ph.GA↗

arXiv2025

ODE-GS: Latent ODEs for Dynamic Scene Extrapolation with 3D Gaussian Splatting

We introduce ODE-GS, a novel approach that integrates 3D Gaussian Splatting with latent neural ordinary differential equations (ODEs) to enable future extrapolation of dynamic 3D scenes. Unlike existing dynamic scene reconstruction methods, which rely on time-conditioned deformation networks and are limited to interpolation within a fixed time window, ODE-GS eliminates timestamp dependency by modeling Gaussian parameter trajectories as continuous-time latent dynamics. Our approach first learns an interpolation model to generate accurate Gaussian trajectories within the observed window, then trains a Transformer encoder to aggregate past trajectories into a latent state evolved via a neural ODE. Finally, numerical integration produces smooth, physically plausible future Gaussian trajectories, enabling rendering at arbitrary future timestamps. On the D-NeRF, NVFi, and HyperNeRF benchmarks, ODE-GS achieves state-of-the-art extrapolation performance, improving metrics by 19.8% compared to leading baselines, demonstrating its ability to accurately represent and predict 3D scene dynamics.

cs.GR↗

arXiv2025

Hybrid Brain-Machine Interface: Integrating EEG and EMG for Reduced Physical Demand

We present a hybrid brain-machine interface (BMI) that integrates steady-state visually evoked potential (SSVEP)-based EEG and facial EMG to improve multimodal control and mitigate fatigue in assistive applications. Traditional BMIs relying solely on EEG or EMG suffer from inherent limitations; EEG-based control requires sustained visual focus, leading to cognitive fatigue, while EMG-based control induces muscular fatigue over time. Our system dynamically alternates between EEG and EMG inputs, using EEG to detect SSVEP signals at 9.75 Hz and 14.25 Hz and EMG from cheek and neck muscles to optimize control based on task demands. In a virtual turtle navigation task, the hybrid system achieved task completion times comparable to an EMG-only approach, while 90% of users reported reduced or equal physical demand. These findings demonstrate that multimodal BMI systems can enhance usability, reduce strain, and improve long-term adherence in assistive technologies.

q-bio.NC↗

arXiv2025

HOMER: Homography-Based Efficient Multi-view 3D Object Removal

3D object removal is an important sub-task in 3D scene editing, with broad applications in scene understanding, augmented reality, and robotics. However, existing methods struggle to achieve a desirable balance among consistency, usability, and computational efficiency in multi-view settings. These limitations are primarily due to unintuitive user interaction in the source view, inefficient multi-view object mask generation, computationally expensive inpainting procedures, and a lack of applicability across different radiance field representations. To address these challenges, we propose a novel pipeline that improves the quality and efficiency of multi-view object mask generation and inpainting. Our method introduces an intuitive region-based interaction mechanism in the source view and eliminates the need for camera poses or extra model training. Our lightweight HoMM module is employed to achieve high-quality multi-view mask propagation with enhanced efficiency. In the inpainting stage, we further reduce computational costs by performing inpainting only on selected key views and propagating the results to other views via homography-based mapping. Our pipeline is compatible with a variety of radiance field frameworks, including NeRF and 3D Gaussian Splatting, demonstrating improved generalizability and practicality in real-world scenarios. Additionally, we present a new 3D multi-object removal dataset with greater object diversity and viewpoint variation than existing datasets. Experiments on public benchmarks and our proposed dataset show that our method achieves state-of-the-art performance while reducing runtime to one-fifth of that required by leading baselines.

cs.CV↗

arXiv2024

Iris: Integrating Language into Diffusion-based Monocular Depth Estimation

Traditional monocular depth estimation suffers from inherent ambiguity and visual nuisances. We demonstrate that language can enhance monocular depth estimation by providing an additional condition (rather than images alone) aligned with plausible 3D scenes, thereby reducing the solution space for depth estimation. This conditional distribution is learned during the text-to-image pre-training of diffusion models. To generate images under various viewpoints and layouts that precisely reflect textual descriptions, the model implicitly models object sizes, shapes, and scales, their spatial relationships, and the overall scene structure. In this paper, Iris, we investigate the benefits of our strategy to integrate text descriptions into training and inference of diffusion-based depth estimation models. We experiment with three different diffusion-based monocular depth estimators (Marigold, Lotus, and E2E-FT) and their variants. By training on HyperSim and Virtual KITTI, and evaluating on NYUv2, KITTI, ETH3D, ScanNet, and DIODE, we find that our strategy improves the overall monocular depth estimation accuracy, especially in small areas. It also improves the model's depth perception of specific regions described in the text. We find that by providing more details in the text, the depth prediction can be iteratively refined. Simultaneously, we find that language can act as a constraint to accelerate the convergence of both training and the inference diffusion trajectory. Code and generated text data will be released upon acceptance.

cs.CV↗

arXiv2024

Comparing Fairness of Generative Mobility Models

This work examines the fairness of generative mobility models, addressing the often overlooked dimension of equity in model performance across geographic regions. Predictive models built on crowd flow data are instrumental in understanding urban structures and movement patterns; however, they risk embedding biases, particularly in spatiotemporal contexts where model performance may reflect and reinforce existing inequities tied to geographic distribution. We propose a novel framework for assessing fairness by measuring the utility and equity of generated traces. Utility is assessed via the Common Part of Commuters (CPC), a similarity metric comparing generated and real mobility flows, while fairness is evaluated using demographic parity. By reformulating demographic parity to reflect the difference in CPC distribution between two groups, our analysis reveals disparities in how various models encode biases present in the underlying data. We utilized four models (Gravity, Radiation, Deep Gravity, and Non-linear Gravity) and our results indicate that traditional gravity and radiation models produce fairer outcomes, although Deep Gravity achieves higher CPC. This disparity underscores a trade-off between model accuracy and equity, with the feature-rich Deep Gravity model amplifying pre-existing biases in community representations. Our findings emphasize the importance of integrating fairness metrics in mobility modeling to avoid perpetuating inequities.

cs.LG↗

arXiv2024

RSA: Resolving Scale Ambiguities in Monocular Depth Estimators through Language Descriptions

We propose a method for metric-scale monocular depth estimation. Inferring depth from a single image is an ill-posed problem due to the loss of scale from perspective projection during the image formation process. Any scale chosen is a bias, typically stemming from training on a dataset; hence, existing works have instead opted to use relative (normalized, inverse) depth. Our goal is to recover metric-scaled depth maps through a linear transformation. The crux of our method lies in the observation that certain objects (e.g., cars, trees, street signs) are typically found or associated with certain types of scenes (e.g., outdoor). We explore whether language descriptions can be used to transform relative depth predictions to those in metric scale. Our method, RSA, takes as input a text caption describing objects present in an image and outputs the parameters of a linear transformation which can be applied globally to a relative depth map to yield metric-scaled depth predictions. We demonstrate our method on recent general-purpose monocular depth models on indoors (NYUv2, VOID) and outdoors (KITTI). When trained on multiple datasets, RSA can serve as a general alignment module in zero-shot settings. Our method improves over common practices in aligning relative to metric depth and results in predictions that are comparable to an upper bound of fitting relative depth to ground truth via a linear transformation.

cs.CV↗