arXiv subjects

Jonathan Liu

Publications and source records attributed to Jonathan Liu.

11 recordsLinked to original sources

arXiv2026

DexDirect: Direct Kinesthetic Arm Guidance for Efficient Dexterous Demonstration Collection

Scalable collection of dexterous manipulation demonstrations remains a major bottleneck for robot learning. High-fidelity interfaces often require costly hardware and extensive setup, while low-setup, low cost alternatives tend to provide less precise control and impose greater cognitive workload on operators. We present DexDirect, a direct kinesthetic arm guidance for efficient dexterous demonstration collection. The operator drags a 6-DoF gravity-compensated robot arm directly by a handle, while a single webcam retargets operator's other hand onto a 16 joints 13-DoF dexterous robot hand. User studies suggest DexDirect collects 17.2x and 3.2x more successful demonstrations compared to purely vision (AnyTeleop) and pose-tracking (TeleDex) baselines. An adapted NASA-TLX shows DexDirect greatly reduces mental demand, effort, and frustration, despite raising physical demand. A diffusion policy trained on DexDirect demonstrations reaches a 90% success rate on a cube pick-and-place task. These results suggest that direct kinesthetic arm guidance combined with vision-based hand retargeting provides an efficient low-setup and scalable interface for collecting dexterous manipulation demonstrations

cs.RO↗

arXiv2026

The Bi-UFS Positive Conjecture for algebraic semidomains

A semidomain is called bi-UFS if both its additive monoid and its nonzero multiplicative monoid are unique factorization monoids. The Bi-UFS Positive Conjecture predicts that the only positive semidomain with this property is the nonnegative integers. We prove this conjecture for finitely generated algebraic positive semidomains. In the cyclic case, we show that for every positive algebraic number $\alpha$, the semidomain $\mathbb{N}_0[\alpha]$ is bi-UFS if and only if $\alpha \in \mathbb{N}$, equivalently $\mathbb{N}_0[\alpha]=\mathbb{N}_0$. The proof separates the quadratic case, where an analysis of the least additive atom larger than $1$ leaves only the examples $\mathbb{N}_0[\sqrt 2]$ and $\mathbb{N}_0[(1+\sqrt 5)/2]$ to exclude, from the higher-degree case, where explicit multiplicative identities force the minimal polynomial into impossible forms. We then give a Perron-Frobenius argument showing that if $\alpha_1,\ldots,\alpha_n$ are positive algebraic numbers and $\mathbb{N}_0[\alpha_1,\ldots,\alpha_n]$ is bi-UFS then this semidomain is $\mathbb{N}_0$. Finally, we prove a reduction theorem for complex semidomains: every bi-UFS subsemidomain of $\mathbb{C}$ with finitely many additive atoms admits an isomorphic realization as a positive semidomain. Consequently, every finitely generated algebraic bi-UFS semidomain over $\mathbb{C}$ is isomorphic to $\mathbb{N}_0$.

math.AC↗

arXiv2026

AGRI-Fidelity: Evaluating the Reliability of Listenable Explanations for Poultry Disease Detection

Existing XAI metrics measure faithfulness for a single model, ignoring model multiplicity where near-optimal classifiers rely on different or spurious acoustic cues. In noisy farm environments, stationary artifacts such as ventilation noise can produce explanations that are faithful yet unreliable, as masking-based metrics fail to penalize redundant shortcuts. We propose AGRI-Fidelity, a reliability-oriented evaluation framework for listenable explanations in poultry disease detection without spatial ground truth. The method combines cross-model consensus with cyclic temporal permutation to construct null distributions and compute a False Discovery Rate (FDR), suppressing stationary artifacts while preserving time-localized bioacoustic markers. Across real and controlled datasets, AGRI-Fidelity effectively provides reliability-aware discrimination for all data points versus masking-based metrics.

cs.LG↗

arXiv2026

Continuous Diffusion Transformers for Designing Synthetic Regulatory Elements

We present a parameter-efficient Diffusion Transformer (DiT) for generating 200bp cell-type-specific regulatory DNA sequences. By replacing the U-Net backbone of DNA-Diffusion with a transformer denoiser equipped with a 2D CNN input encoder, our model matches the U-Net's best validation loss in 13 epochs (60$\times$ fewer) and converges 39% lower, while reducing memorization from 5.3% to 1.7% of generated sequences aligning to training data via BLAT. Ablations show the CNN encoder is essential: without it, validation loss increases 70% regardless of positional embedding choice. We further apply DDPO finetuning using Enformer as a reward model, achieving a 38$\times$ improvement in predicted regulatory activity. Cross-validation against DRAKES on an independent prediction task confirms that improvements reflect genuine regulatory signal rather than reward model overfitting.

cs.LG↗

arXiv2025

Demo: Statistically Significant Results On Biases and Errors of LLMs Do Not Guarantee Generalizable Results

Recent research has shown that hallucinations, omissions, and biases are prevalent in everyday use-cases of LLMs. However, chatbots used in medical contexts must provide consistent advice in situations where non-medical factors are involved, such as when demographic information is present. In order to understand the conditions under which medical chatbots fail to perform as expected, we develop an infrastructure that 1) automatically generates queries to probe LLMs and 2) evaluates answers to these queries using multiple LLM-as-a-judge setups and prompts. For 1), our prompt creation pipeline samples the space of patient demographics, histories, disorders, and writing styles to create realistic questions that we subsequently use to prompt LLMs. In 2), our evaluation pipeline provides hallucination and omission detection using LLM-as-a-judge as well as agentic workflows, in addition to LLM-as-a-judge treatment category detectors. As a baseline study, we perform two case studies on inter-LLM agreement and the impact of varying the answering and evaluation LLMs. We find that LLM annotators exhibit low agreement scores (average Cohen's Kappa $\kappa=0.118$), and only specific (answering, evaluation) LLM pairs yield statistically significant differences across writing styles, genders, and races. We recommend that studies using LLM evaluation use multiple LLMs as evaluators in order to avoid arriving at statistically significant but non-generalizable results, particularly in the absence of ground-truth data. We also suggest publishing inter-LLM agreement metrics for transparency. Our code and dataset are available here: https://github.com/BBN-E/medic-neurips-2025-demo.

cs.CL↗

arXiv2025

MobiVerse: Scaling Urban Mobility Simulation with Hybrid Lightweight Domain-Specific Generator and Large Language Models

Understanding and modeling human mobility patterns is crucial for effective transportation planning and urban development. Despite significant advances in mobility research, there remains a critical gap in simulation platforms that allow for algorithm development, policy implementation, and comprehensive evaluation at scale. Traditional activity-based models require extensive data collection and manual calibration, machine learning approaches struggle with adaptation to dynamic conditions, and treding agent-based Large Language Models (LLMs) implementations face computational constraints with large-scale simulations. To address these challenges, we propose MobiVerse, a hybrid framework leverages the efficiency of lightweight domain-specific generator for generating base activity chains with the adaptability of LLMs for context-aware modifications. A case study was conducted in Westwood, Los Angeles, where we efficiently generated and dynamically adjusted schedules for the whole population of approximately 53,000 agents on a standard PC. Our experiments demonstrate that MobiVerse successfully enables agents to respond to environmental feedback, including road closures, large gathering events like football games, and congestion, through our hybrid framework. Its modular design facilitates testing various mobility algorithms at both transportation system and agent levels. Results show our approach maintains computational efficiency while enhancing behavioral realism. MobiVerse bridges the gap in mobility simulation by providing a customizable platform for mobility systems planning and operations with benchmark algorithms. Code and videos are available at https://github.com/ucla-mobility/MobiVerse.

cs.AI↗

arXiv2025

Procedural Generation of Articulated Simulation-Ready Assets

We introduce Infinigen-Articulated, a toolkit for generating realistic, procedurally generated articulated assets for robotics simulation. We include procedural generators for 18 common articulated object categories along with high-level utilities for use creating custom articulated assets in Blender. We also provide an export pipeline to integrate the resulting assets along with their physical properties into common robotics simulators. Experiments demonstrate that assets sampled from these generators are effective for movable object segmentation, training generalizable reinforcement learning policies, and sim-to-real transfer of imitation learning policies.

cs.RO↗

arXiv2024

Teaching Algorithm Design: A Literature Review

Algorithm design is a vital skill developed in most undergraduate Computer Science (CS) programs, but few research studies focus on pedagogy related to algorithms coursework. To understand the work that has been done in the area, we present a systematic survey and literature review of CS Education studies. We search for research that is both related to algorithm design and evaluated on undergraduate-level students. Across all papers in the ACM Digital Library prior to August 2023, we only find 94 such papers. We first classify these papers by topic, evaluation metric, evaluation methods, and intervention target. Through our classification, we find a broad sparsity of papers which indicates that many open questions remain about teaching algorithm design, with each algorithm topic only being discussed in between 0 and 10 papers. We also note the need for papers using rigorous research methods, as only 38 out of 88 papers presenting quantitative data use statistical tests, and only 15 out of 45 papers presenting qualitative data use a coding scheme. Only 17 papers report controlled trials. We then synthesize the results of the existing literature to give insights into what the corpus reveals about how we should teach algorithms. Much of the literature explores implementing well-established practices, such as active learning or automated assessment, in the algorithms classroom. However, there are algorithms-specific results as well: a number of papers find that students may under-utilize certain algorithmic design techniques, and studies describe a variety of ways to select algorithms problems that increase student engagement and learning. The results we present, along with the publicly available set of papers collected, provide a detailed representation of the current corpus of CS Education work related to algorithm design and can orient further research in the area.

cs.DS↗

arXiv2021

A projection-domain low-count quantitative SPECT method for alpha-particle emitting radiopharmaceutical therapy

Single-photon emission computed tomography (SPECT) provides a mechanism to estimate regional isotope uptake in lesions and at-risk organs after administration of {\alpha}-particle-emitting radiopharmaceutical therapies ({\alpha}-RPTs). However, this estimation task is challenging due to the complex emission spectra, the very low number of detected counts, the impact of stray-radiation-related noise at these low counts, and the multiple image-degrading processes in SPECT. The conventional reconstruction-based quantification methods are observed to be erroneous for {\alpha}-RPT SPECT. To address these challenges, we developed a low-count quantitative SPECT (LC-QSPECT) method that directly estimates the regional activity uptake from the projection data, compensates for stray-radiation-related noise, and for the radioisotope and SPECT physics. The method was validated in the context of three-dimensional SPECT with 223 Ra. Validation was performed using both realistic simulation studies, including a virtual clinical trial, and synthetic and anthropomorphic physical-phantom studies. Across all studies, the LC-QSPECT method yielded reliable regional-uptake estimates and outperformed the conventional ordered subset expectation maximization (OSEM)-based reconstruction and geometric transfer matrix (GTM)-based post-reconstruction partial-volume compensation methods. Further, the method yielded reliable uptake across different lesion sizes, contrasts, and different levels of intra-lesion heterogeneity. Additionally, the variance of the estimated uptake approached the Cram\'e-Rao bound-defined theoretical limit.

physics.med-ph↗

arXiv2021

Heated gas bubbles enrich, crystallize, dry, phosphorylate and encapsulate prebiotic molecules

Non-equilibrium conditions must have been crucial for the assembly of the first informational polymers of early life, but supporting their formation and continuous enrichment in a long-lasting environment. Here, we explore how gas bubbles in water subjected to a thermal gradient, a likely scenario within crustal mafic rocks on the early Earth, drive a complex, continuous enrichment of prebiotic molecules. NRA precursors, monomers, active ribozymes, oligonucleotides and lipids are shown to (1) cycle between dry and wet states, enabling the central step of RNA phosphorylation, (2) accumulate at the gas-water interface to drastically increase ribozymatic activity, (3) condense into hydrogels, (4) form pure crystals and (5) encapsulate into protecting vesicle aggregates that subsequently undergo fission. These effects occur within less than 30 min. The findings unite, in one location, the physical conditions that were crucial for the chemical emergence of biopolymers. They suggest that heated microbubbles could have hosted the first cycles of molecular evolution.

cond-mat.soft↗

arXiv2019

Algorithmic Discrepancy Minimization

This report will be a literature review on a result in algorithmic discrepancy theory. We will begin by providing a quick overview on discrepancy theory and some major results in the field, and then focus on an important result by Shachar Lovett and Raghu Meka. We restate the main algorithm and ideas of the paper, and rewrite proofs for some of the major results in the paper.

cs.DS↗