Publications
Our teams aspire to make discoveries that impact everyone, and core to our approach is sharing our research and tools to fuel progress in the field
Our teams aspire to make discoveries that impact everyone, and core to our approach is sharing our research and tools to fuel progress in the field
Sort By
1 - 15 of 10923 publications
mmMUSE: An mmWave-based Motion-resilient Universal Speech Enhancement System
Chenming He
Yanyong Zhang
Kai Wang
Dequan Wang
Lingyu Wang
the Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT), ACM (2026) (to appear)
Preview abstract
Voice-based smart systems can greatly enhance user experiences by allowing higher-quality interactions through better voice perception. Speech enhancement can benefit such systems by isolating noise from speech. Recently, integrating millimeter-wave (mmWave) with audio for speech perception has gained increasing attention due to microphones' limitations in noisy environments. However, mmWave-based vocal extraction is severely affected by motion, which disperses vocal signals across ranges and introduces distortions. In this paper, we propose an mmWave-based motion-resilient universal speech enhancement system called mmMUSE, which fuses mmWave and audio signals. To mitigate motion interference, we develop a Doppler-based method for motion-robust vocal signal extraction. Moreover, by introducing the Vocal-Noise-Ratio metric to assess the prominence of vocal signals from mmWave, we achieve real-time voice activity detection that gains 3.81 dB of SISDR in noisy speeches. Additionally, we design a two-stage complex-valued network that includes an attention-based fusion network for cross-modal complementing and a time-frequency masking network for correcting amplitude and phase of speech to isolate noises.
Using mmWave and audio datasets from 46 participants, mmMUSE outperforms the state-of-the-art speech enhancement models, achieving an average SISDR improvement of 3.12 dB. Additionally, mmMUSE achieves SISDR improvements of 16.51 dB, 17.93 dB, 14.93 dB, and 18.95 dB in controlled environments involving intense noise, extensive motion, multiple speakers, and various obstructive materials, respectively. Finally, we evaluate mmMUSE in real-world scenarios including running, public spaces, and driving, maintaining a word error rate (WER) below 10%.
View details
Preview abstract
How many T gates are needed to approximate an arbitrary n-qubit quantum state to within
a given precision ϵ? Improving prior work of Low, Kliuchnikov and Schaeffer, we show that the
optimal asymptotic scaling is Θ(sqrt{2^n log(1/ε)} + log(1/ε)) if we allow an unlimited number of ancilla qubits. We also show that this is the optimal T-count for implementing an arbitrary
diagonal n-qubit unitary to within error ϵ. We describe an application to batched synthesis of
single-qubit unitaries: we can approximate a tensor product of m = O(log log(1/ϵ)) arbitrary
single-qubit unitaries to within error ϵ with the same asymptotic T-count as is required to
approximate just one single-qubit unitary.
View details
Preview abstract
For many practical applications of quantum computing, the slowest and most costly steps involve coherently accessing classical data. We help address this challenge by applying mass production techniques, which can sometimes allow us to perform operations many times in parallel for a cost that is comparable to a single execution[1-3]. We combine existing mass-production results with modern approaches for loading classical data using ``quantum read-only memory.'' We show that quantum mass production techniques offer no benefit when we consider a cost model that focuses purely on the number of non-Clifford gates. However, analyzing the constant factors in a more nuanced cost model, we find that it may be possible to obtain a reduction in cost of an order or magnitude or more for a variety reasonably-sized fault-tolerant quantum algorithms. We present several applications of quantum mass-production techniques beyond naive parallelization, including a strategy for reducing the cost of serial calls to the same data loading step.
View details
Preview abstract
AI coding assistants are rapidly becoming integral to modern software development. A key challenge in this space is the continual need to migrate and modernize codebases in response to evolving software ecosystems. Traditionally, such migrations have relied on rule-based systems and human intervention. With the advent of powerful large language models (LLMs), AI-driven agentic frameworks offer a promising alternative—but their effectiveness remains underexplored. In this paper, we introduce FreshBrew, a novel benchmark for evaluating AI-based agentic frameworks on project-level Java migrations. We benchmark several such frameworks, powered by state-of-the-art LLMs, and compare their performance against established rule-based tools. Our evaluation of AI agents on this benchmark of 228 repositories shows that the top-performing model, Gemini 2.5 Flash, can successfully migrate 56.5% of projects to JDK 17. Our empirical analysis reveals novel insights into the critical strengths and limitations of current agentic approaches, offering actionable insights into their real-world applicability. By releasing FreshBrew publicly upon acceptance, we aim to facilitate rigorous, reproducible evaluation and catalyze progress in AI-driven codebase modernization.
View details
CrossCheck: Input Validation for WAN Control Systems
Bharath Modhipalli
Rishabh Iyer
Isaac Keslassy
Sylvia Ratnasamy
Networked Systems Design and Implementation (NSDI) (2026) (to appear)
Preview abstract
We present CrossCheck, a system that validates inputs to the Software-Defined Networking (SDN) controller in a Wide Area Network (WAN). By detecting incorrect inputs—often stemming from bugs in the SDN control infrastructure—CrossCheck alerts operators before they trigger network outages.
Our analysis at a large-scale WAN operator identifies invalid inputs as a leading cause of major outages, and we show how CrossCheck would have prevented those incidents. We deployed CrossCheck as a shadow validation system for four weeks in a production WAN, during which it accurately detected the single incident of invalid inputs that occurred while sustaining a 0% false positive rate under normal operation, hence imposing little additional burden on operators. In addition, we show through simulation that CrossCheck reliably detects a wide range of invalid inputs (e.g., detecting demand perturbations as small as 5% with 100% accuracy) and maintains a near-zero false positive rate for realistic levels of noisy, missing, or buggy telemetry data (e.g., sustaining zero false positives with up to 30% of corrupted telemetry data).
View details
The FLuid Allocation of Surface Code Qubits (FLASQ) Cost Model for Early Fault-Tolerant Quantum Algorithms
Bill Huggins
Amanda Xu
Matthew Harrigan
Christopher Kang
Guang Hao Low
Austin Fowler
arXiv:2511.08508 (2025)
Preview abstract
Holistic resource estimates are essential for guiding the development of fault-tolerant quantum algorithms and the computers they will run on. This is particularly true when we focus on highly-constrained early fault-tolerant devices. Many attempts to optimize algorithms for early fault-tolerance focus on simple metrics, such as the circuit depth or T-count. These metrics fail to capture critical overheads, such as the spacetime cost of Clifford operations and routing, or miss they key optimizations. We propose the FLuid Allocation of Surface code Qubits (FLASQ) cost model, tailored for architectures that use a two-dimensional lattice of qubits to implement the two-dimensional surface code. FLASQ abstracts away the complexity of routing by assuming that ancilla space and time can be fluidly rearranged, allowing for the tractable estimation of spacetime volume while still capturing important details neglected by simpler approaches. At the same time, it enforces constraints imposed by the circuit's measurement depth and the processor's reaction time. We apply FLASQ to analyze the cost of a standard two-dimensional lattice model simulation, finding that modern advances (such as magic state cultivation and the combination of quantum error correction and mitigation) reduce both the time and space required for this task by an order of magnitude compared with previous estimates. We also analyze the Hamming weight phasing approach to synthesizing parallel rotations, revealing that despite its low T-count, the overhead from imposing a 2D layout and from its use of additional ancilla qubits will make it challenging to benefit from in early fault-tolerance. We hope that the FLASQ cost model will help to better align early fault-tolerant algorithmic design with actual hardware realization costs without demanding excessive knowledge of quantum error correction from quantum algorithmists.
View details
I know what I don't know: improving model cascades through confidence tuning
Stephan Rabanser
Nathalie Rauschmayr
Petra Poklukar
Congchao Wang
2025
Preview abstract
Large-scale machine learning models deliver strong performance across a wide range of tasks but come with significant computational and resource constraints. To mitigate these challenges, local smaller models are often deployed alongside larger models, relying on routing and deferral mechanisms to offload complex tasks. However, existing approaches inadequately balance the capabilities of these models, often resulting in unnecessary deferrals or sub-optimal resource usage. In this work we introduce a novel loss function called Gatekeeper for calibrating smaller models in cascade setups. Our approach fine-tunes the smaller model to confidently handle tasks it can perform correctly while deferring complex tasks to the larger model. Moreover, it incorporates a mechanism for managing the trade-off between model performance and deferral accuracy, and is broadly applicable across various tasks and domains without any architectural changes. We evaluated our method on encoder-only, decoder-only, and encoder-decoder architectures. Experiments across image classification, language modeling, and vision-language tasks show that our approach substantially improves deferral performance.
View details
Preview abstract
Augmenting LLMs with context leads to improved performance across many applications. Despite much research on Retrieval Augmented Generation (RAG) systems, an open question is whether errors arise because LLMs fail to utilize the context from retrieval or the context itself is insufficient to answer the query. To shed light on this, we develop a new notion of sufficient context, along with a way to classify instances that have enough information to answer the query. We then use sufficient context to analyze several models and datasets. By stratifying errors based on context sufficiency, we find that proprietary LLMs (Gemini, GPT, Claude) excel at answering queries when the context is sufficient, but often output incorrect answers instead of abstaining when the context is not. On the other hand, open-source LLMs (Llama, Mistral, Gemma) hallucinate or abstain often, even with sufficient context. We further categorize cases when the context is useful, and improves accuracy, even though it does not fully answer the query and the model errs without the context. Building on our findings, we explore ways to reduce hallucinations in RAG systems, including a new selective generation method that leverages sufficient context information for guided abstention. Our method improves the fraction of correct answers among times where the model responds by 2--10% for Gemini, GPT, and Gemma.
View details
Preview abstract
A growing body of research has demonstrated that the behavior of large language models can be effectively controlled at inference time by directly modifying their internal states, either through vector additions to their activations or through updates to their weight matrices. These techniques, while powerful, are often guided by empirical heuristics, such as deriving ``steering vectors'' from the average activations of contrastive prompts. This work provides a theoretical foundation for these interventions, explaining how they emerge from the fundamental computations of the transformer architecture. Building on the recent finding that a prompt's influence can be mathematically mapped to implicit weight updates Dherin et al. (2025), we generalize this theory to deep, multi-block transformers. We show how the information contained in any chunk of a user prompt is represented and composed internally through virtual weight vectors and virtual weight matrices. We then derive a principled method for condensing this information into token-independent thought vectors and thought matrices. These constructs provide a theoretical explanation for existing vector- and matrix-based model editing techniques and offer a direct, computationally-grounded method for transforming textual input into reusable weight updates.
View details
Preview abstract
Recently, Chevignard et al proposed a way to factor $n$ bit RSA integers using only $(0.5 + \epsilon)n$ logical qubits.
In this paper, I streamline Chevignard's algorithm and estimate its physical cost accounting for the overhead of error correction.
I reduce its Toffoli count by more than 100x, and show that this implies a 2048 bit RSA integer could be factored in less than a week using less than one million noisy qubits (compared to 20 million in Gidney+Eker{\aa} 2019).
I make the same assumptions as in Gidney+Eker{\aa} 2019: a square grid of qubits with nearest neighbor connections, a gate error rate of $0.1\%$, a surface code cycle time of 1 microsecond, and a control system reaction time of $10$ microseconds.
View details
Preview abstract
Automating data visualization from natural language is crucial for data science, yet current systems struggle with complex, multi-file datasets and iterative refinement. Existing approaches, including simple single- or multi-agent systems, often oversimplify the task, focusing on initial query parsing while failing to robustly manage data complexity, code errors, or final visualization quality. In this paper, we reframe this challenge as a collaborative multi-agent problem. We introduce CoDA, a multi-agent system that employs specialized LLM agents for metadata analysis, task planning, code generation, and iterative reflection. We formalize this pipeline, demonstrating how metadata-focused analysis bypasses token limits and quality-driven refinement ensures robustness. Extensive evaluations show CoDA achieves substantial accuracy gains, outperforming competitive baselines by up to 49.0%. This work advocates that future visualization automation should evolve from isolated code generation to integrated, collaborative agentic workflows.
View details
How to deal w___ missing input data
Martin Gauch
Frederik Kratzert
Daniel Klotz
Hydrology and Earth System Sciences, 29 (2025), pp. 6221-6235
Preview abstract
Deep learning hydrologic models have made their way from research to applications. More and more national hydrometeorological agencies, hydro power operators, and engineering consulting companies are building Long Short-Term Memory (LSTM) models for operational use cases. All of these efforts come across similar sets of challenges – challenges that are different from those in controlled scientific studies. In this paper, we tackle one of these issues: how to deal with missing input data? Operational systems depend on the real-time availability of various data products – most notably, meteorological forcings. The more external dependencies a model has, however, the more likely it is to experience an outage in one of them. We introduce and compare three different solutions that can generate predictions even when some of the meteorological input data do not arrive in time, or not arrive at all: First, input replacing, which imputes missing values with a fixed number; second, masked mean, which averages embeddings of the forcings that are available at a given time step; third, attention, a generalization of the masked mean mechanism that dynamically weights the embeddings. We compare the approaches in different missing data scenarios and find that, by a small margin, the masked mean approach tends to perform best.
View details
Preview abstract
Users of routing services like Apple Maps, Google Maps, and Waze frequently wonder why a given route is proposed. This question particularly arises when dynamic conditions like traffic and road closures cause unusual routes to be proposed. While many such dynamic conditions may exist in a road network at any time, only a small fraction of those conditions are typically relevant to a given user's route. In this work, we give a simple algorithm that identifies a small set of traffic-laden road segments that answer the following question: Which traffic conditions cause a particular shortest traffic-aware route to differ from the shortest traffic-free route? We theoretically and experimentally show that our algorithm generates small and interpretable answers to this question.
View details
HEART: Emotionally-driven test-time scaling of Language Models
Souradip Chakraborty
Gabriela Pinto
2025
Preview abstract
Test-time scaling has shown considerable success in improving the performance of language models on complex reasoning tasks without requiring fine-tuning. However, current strategies, such as self-reflection or ensembling, primarily focus on logical or structural refinement. They do not leverage the guiding potential of affective feedback. Inspired by psychological research showing that emotions can modulate cognitive performance, we introduce HEART--a novel framework that uses emotionally-driven prompts for iterative self-correction. HEART provides feedback on a models' incorrect response using a curated set of concise, emotionally charged phrases based on Paul Ekman's six basic emotions. By systematically varying the emotional tone of the feedback across iterations, our method guides the model to escape flawed reasoning paths and explore more promising alternatives. We evaluate our framework on challenging reasoning benchmarks including OlympiadBench, Humanity's Last Exam, and SimpleQA. Across these benchmarks, our approach delivers significantly deeper reasoning which leads to consistent and significant increase in accuracy compared to existing prompting methods. Crucially, these gains are observed across a diverse range of model architectures, demonstrating the broad applicability of our technique. Overall, our findings suggest that the next frontier in machine reasoning may lie not just in refining logic, but also in understanding and leveraging the 'HEART' of the models.
View details
Preview abstract
We study the allocation of indivisible goods under conflicting constraints, represented by a graph. In this framework, vertices correspond to goods and edges correspond to conflicts between a pair of goods. Each agent is allocated an independent set in the graph. In a recent work of [Kumar et al., 2024], it was shown that a maximal EF1 allocation exists for interval graphs and two agents with monotone valuations. We significantly extend this result by establishing that a maximal EF1 allocation exists for *any graph* when the two agents have monotone valuations. To compute such an allocation, we present a polynomial-time algorithm for additive valuations as well as a pseudo-polynomial time algorithm for monotone valuations. Moreover, we complement our findings by providing a counter example demonstrating a maximal EF1 allocation may not exist for three agents with monotone valuations. Additionally, we establish NP-hardness of determining the existence of such allocations for every fixed number n of agents.
View details
