arXiv subjects

Cindy Wang

Publications and source records attributed to Cindy Wang.

11 recordsLinked to original sources

arXiv2026

OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning

We introduce OfficeQA Pro, a benchmark for evaluating AI agents on grounded, multi-document reasoning over a large and heterogeneous document corpus. The corpus consists of U.S. Treasury Bulletins spanning nearly 100 years, comprising 89,000 pages and over 26 million numerical values. OfficeQA Pro consists of 133 questions that require precise document parsing, retrieval, and analytical reasoning across both unstructured text and tabular data. Frontier LLMs including Claude Opus 4.6, GPT-5.4, and Gemini 3.1 Pro Preview achieve less than 5% accuracy on OfficeQA Pro when relying on parametric knowledge, and less than 12% with additional access to the web. When provided directly with the document corpus, frontier agents still struggle on over half of questions, scoring 34.1% on average. We find that providing agents with a structured document representation produced by Databricks' ai_parse_document yields a 16.1% average relative performance gain across agents. We conduct additional ablations to study the effects of model selection, table representation, retrieval strategy, and test-time scaling on performance. Despite these improvements, significant headroom remains before agents can be considered reliable at enterprise-grade grounded reasoning.

cs.AI↗

arXiv2026

KARL: Knowledge Agents via Reinforcement Learning

We present a system for training enterprise search agents via reinforcement learning that achieves state-of-the-art performance across a diverse suite of hard-to-verify agentic search tasks. Our work makes four core contributions. First, we introduce KARLBench, a multi-capability evaluation suite spanning six distinct search regimes, including constraint-driven entity search, cross-document report synthesis, tabular numerical reasoning, exhaustive entity retrieval, procedural reasoning over technical documentation, and fact aggregation over internal enterprise notes. Second, we show that models trained across heterogeneous search behaviors generalize substantially better than those optimized for any single benchmark. Third, we develop an agentic synthesis pipeline that employs long-horizon reasoning and tool use to generate diverse, grounded, and high-quality training data, with iterative bootstrapping from increasingly capable models. Fourth, we propose a new post-training paradigm based on iterative large-batch off-policy RL that is sample efficient, robust to train-inference engine discrepancies, and naturally extends to multi-task training with out-of-distribution generalization. Compared to Claude 4.6 and GPT 5.2, KARL is Pareto-optimal on KARLBench across cost-quality and latency-quality trade-offs, including tasks that were out-of-distribution during training. With sufficient test-time compute, it surpasses the strongest closed models. These results show that tailored synthetic data in combination with multi-task reinforcement learning enables cost-efficient and high-performing knowledge agents for grounded reasoning.

cs.AI↗

arXiv2026

Bidirectional Human-Robot Communication for Physical Human-Robot Interaction

Effective physical human-robot interaction requires systems that are not only adaptable to user preferences but also transparent about their actions. This paper introduces BRIDGE, a system for bidirectional human-robot communication in physical assistance. Our method allows users to modify a robot's planned trajectory -- position, velocity, and force -- in real time using natural language. We utilize a large language model (LLM) to interpret any trajectory modifications implied by user commands in the context of the planned motion and conversation history. Importantly, our system provides verbal feedback in response to the user, either assuring any resulting changes or posing a clarifying question. We evaluated our method in a user study with 18 older adults across three assistive tasks, comparing BRIDGE to an ablation without verbal feedback and a baseline. Results show that participants successfully used the system to modify trajectories in real time. Moreover, the bidirectional feedback led to significantly higher ratings of interactivity and transparency, demonstrating that the robot's verbal response is critical for a more intuitive user experience. Videos and code can be found on our project website: https://bidir-comm.github.io/

cs.RO↗

arXiv2025

Embodied Tree of Thoughts: Deliberate Manipulation Planning with Embodied World Model

World models have emerged as a pivotal component in robot manipulation planning, enabling agents to predict future environmental states and reason about the consequences of actions before execution. While video-generation models are increasingly adopted, they often lack rigorous physical grounding, leading to hallucinations and a failure to maintain consistency in long-horizon physical constraints. To address these limitations, we propose Embodied Tree of Thoughts (EToT), a novel Real2Sim2Real planning framework that leverages a physics-based interactive digital twin as an embodied world model. EToT formulates manipulation planning as a tree search expanded through two synergistic mechanisms: (1) Priori Branching, which generates diverse candidate execution paths based on semantic and spatial analysis; and (2) Reflective Branching, which utilizes VLMs to diagnose execution failures within the simulator and iteratively refine the planning tree with corrective actions. By grounding high-level reasoning in a physics simulator, our framework ensures that generated plans adhere to rigid-body dynamics and collision constraints. We validate EToT on a suite of short- and long-horizon manipulation tasks, where it consistently outperforms baselines by effectively predicting physical dynamics and adapting to potential failures. Website at https://embodied-tree-of-thoughts.github.io .

cs.RO↗

arXiv2025

Improving structure search with hyperspatial optimization and TETRIS seeding

Advanced structure prediction methods developed over the past decades include an unorthodox strategy of allowing atoms to displace into extra dimensions. A recently implemented global optimization of structures from hyperspace (GOSH) has shown promise in accelerating the identification of global minima on potential energy surfaces defined by simple interatomic models. In this study, we extend the GOSH formalism to more accurate Behler-Parrinello neural network (NN) potentials, make it compatible with efficient local minimization algorithms, and test its performance on nanoparticles and crystalline solids. For clusters modeled with NN potentials, four-dimensional optimization offers fairly modest improvement in navigating geometric relaxation pathways and incurs increased computational cost largely offsetting the benefit, but it provides a significant advantage in facilitating atom swaps in nanoalloys. In comparison, the introduction of a moderate, controlled bias for generating more physically sensible starting configurations, achieved via TETRIS-inspired packing of atomic blocks, has a more direct impact on the efficiency of global structure searches. The benchmarked systems are Lennard-Jones clusters, Au or Cu-Pd-Ag nanoparticles and binary Sn alloys described by NN potentials, and compounds with covalent B or BC frameworks modeled with density functional theory

cond-mat.mtrl-sci↗

arXiv2024

Generation, Distillation and Evaluation of Motivational Interviewing-Style Reflections with a Foundational Language Model

Large Foundational Language Models are capable of performing many tasks at a high level but are difficult to deploy in many applications because of their size and proprietary ownership. Many will be motivated to distill specific capabilities of foundational models into smaller models that can be owned and controlled. In the development of a therapeutic chatbot, we wish to distill a capability known as reflective listening, in which a therapist produces reflections of client speech. These reflections either restate what a client has said, or connect what was said to a relevant observation, idea or guess that encourages and guides the client to continue contemplation. In this paper, we present a method for distilling the generation of reflections from a Foundational Language Model (GPT-4) into smaller models. We first show that GPT-4, using zero-shot prompting, can generate reflections at near 100% success rate, superior to all previous methods. Using reflections generated by GPT-4, we fine-tune different sizes of the GPT-2 family. The GPT-2-small model achieves 83% success on a hold-out test set and the GPT-2 XL achieves 90% success. We also show that GPT-4 can help in the labor-intensive task of evaluating the quality of the distilled models, using it as a zero-shot classifier. Using triple-human review as a guide, the classifier achieves a Cohen-Kappa of 0.66, a substantial inter-rater reliability figure.

cs.CL↗

arXiv2023

Labrador: Exploring the Limits of Masked Language Modeling for Laboratory Data

In this work we introduce Labrador, a pre-trained Transformer model for laboratory data. Labrador and BERT were pre-trained on a corpus of 100 million lab test results from electronic health records (EHRs) and evaluated on various downstream outcome prediction tasks. Both models demonstrate mastery of the pre-training task but neither consistently outperform XGBoost on downstream supervised tasks. Our ablation studies reveal that transfer learning shows limited effectiveness for BERT and achieves marginal success with Labrador. We explore the reasons for the failure of transfer learning and suggest that the data generating process underlying each patient cannot be characterized sufficiently using labs alone, among other factors. We encourage future work to focus on joint modeling of multiple EHR data categories and to include tree-based baselines in their evaluations.

cs.CL↗

arXiv2023

Exploring Embeddings for Measuring Text Relatedness: Unveiling Sentiments and Relationships in Online Comments

After the COVID-19 pandemic caused internet usage to grow by 70%, there has been an increased number of people all across the world using social media. Applications like Twitter, Meta Threads, YouTube, and Reddit have become increasingly pervasive, leaving almost no digital space where public opinion is not expressed. This paper investigates sentiment and semantic relationships among comments across various social media platforms, as well as discusses the importance of shared opinions across these different media platforms, using word embeddings to analyze components in sentences and documents. It allows researchers, politicians, and business representatives to trace a path of shared sentiment among users across the world. This research paper presents multiple approaches that measure the relatedness of text extracted from user comments on these popular online platforms. By leveraging embeddings, which capture semantic relationships between words and help analyze sentiments across the web, we can uncover connections regarding public opinion as a whole. The study utilizes pre-existing datasets from YouTube, Reddit, Twitter, and more. We made use of popular natural language processing models like Bidirectional Encoder Representations from Transformers (BERT) to analyze sentiments and explore relationships between comment embeddings. Additionally, we aim to utilize clustering and Kl-divergence to find semantic relationships within these comment embeddings across various social media platforms. Our analysis will enable a deeper understanding of the interconnectedness of online comments and will investigate the notion of the internet functioning as a large interconnected brain.

cs.CL↗

arXiv2020

High-fidelity superconducting quantum processors via laser-annealing of transmon qubits

Scaling the number of qubits while maintaining high-fidelity quantum gates remains a key challenge for quantum computing. Presently, superconducting quantum processors with >50-qubits are actively available. For such systems, fixed-frequency transmons are attractive due to their long coherence and noise immunity. However, scaling fixed-frequency architectures proves challenging due to precise relative frequency requirements. Here we employ laser annealing to selectively tune transmon qubits into desired frequency patterns. Statistics over hundreds of annealed qubits demonstrate an empirical tuning precision of 18.5 MHz, with no measurable impact on qubit coherence. We quantify gate error statistics on a tuned 65-qubit processor, with median two-qubit gate fidelity of 98.7%. Baseline tuning statistics yield a frequency-equivalent resistance precision of 4.7 MHz, sufficient for high-yield scaling beyond 1000-qubit levels. Moving forward, we anticipate selective laser annealing to play a central role in scaling fixed-frequency architectures.

quant-ph↗

arXiv2020

Demonstration of quantum volume 64 on a superconducting quantum computing system

We improve the quality of quantum circuits on superconducting quantum computing systems, as measured by the quantum volume, with a combination of dynamical decoupling, compiler optimizations, shorter two-qubit gates, and excited state promoted readout. This result shows that the path to larger quantum volume systems requires the simultaneous increase of coherence, control gate fidelities, measurement fidelities, and smarter software which takes into account hardware details, thereby demonstrating the need to continue to co-design the software and hardware stack for the foreseeable future.

quant-ph↗

arXiv2019

Identifying MgII Narrow Absorption Lines with Deep Learning

Metal absorption line systems in distant quasar spectra probe of the history of gas content in the universe. The MgII $λλ$ 2796, 2803 doublet is one of the most important absorption lines since it is a proxy of the star formation rate and a tracer of the cold gas associated with high redshift galaxies. Machine learning algorithms have been used to detect absorption lines systems in large sky surveys, such as Principle Component Analysis (PCA), Gaussian Process (GP) and decision trees. A very powerful algorithm in the field of machine learning called deep neural networks, or '' deep learning'' is a new structure of neural network that automatically extracts semantic features from raw data and represents them at a high level. In this paper, we apply a deep convolutional neural network for absorption line detection. We use the previously published DR7 MgII catalog (Zhu et al. 2013) as the training and validation sample and the DR12 MgII catalog as the test set. Our deep learning algorithm is capable of detecting MgII absorption lines with an accuracy of $\sim$94% . It takes only $\sim 9$ seconds to analyze $\sim$ 50000 quasar spectra with our deep neural network, which is ten thousand times faster than traditional methods, while preserving high accuracy with little human interference. Our study shows that Mg II absorption line detection accuracy of a deep neutral network model strongly depends on the filter size in the filter layer of the neural network, and the best results are obtained when the filter size closely matches the absorption feature size.

astro-ph.GA↗