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 10793 publications
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
MaRVL-QA: A Benchmark for Mathematical Reasoning over Visual Landscapes
Nilay Pande
Sahiti Yerramilli
Jayant Tamarapalli
Rynaa Grover
(2025)
Preview abstract
A key frontier for Multimodal Large Language Models (MLLMs) is the ability to perform deep mathematical and spatial reasoning directly from images, moving beyond their established success in semantic description. Mathematical surface plots provide a rigorous testbed for this capability, as they isolate the task of reasoning from the semantic noise common in natural images. To measure progress on this frontier, we introduce MaRVL (Mathematical Reasoning over Visual Landscapes), a new benchmark designed to quantitatively evaluate these core reasoning skills. The benchmark comprises two novel tasks: Topological Counting, identifying and enumerating features like local maxima; and Transformation Recognition, recognizing applied geometric transformations. Generated from a curated library of functions with rigorous ambiguity filtering, our evaluation on MaRVL reveals that even state-of-the-art MLLMs struggle significantly, often resorting to superficial heuristics instead of robust spatial reasoning. MaRVL provides a challenging new tool for the research community to measure progress, expose model limitations, and guide the development of MLLMs with more profound reasoning abilities.
View details
Preview abstract
We study an online finite-horizon Markov Decision Processes with adversarially changing loss and aggregate bandit feedback (a.k.a full-bandit). Under this type of feedback,
the agent observes only the total loss incurred over the entire trajectory, rather than the individual losses at each intermediate step within the trajectory. We introduce the first Policy Optimization algorithms for this setting.
In the known-dynamics case, we achieve the first \textit{optimal} regret bound of $\tilde \Theta(H^2\sqrt{SAK})$, where $K$ is the number of episodes, $H$ is the horizon, $S$ is the number of states, and $A$ is the number of actions of the MDP.
In the unknown dynamics case we establish regret bound of $\tilde O(H^3 S \sqrt{AK})$, significantly improving the best known result by a factor of $H^2 S^5 A^2$.
View details
Synthetic Dialogue Generation for Interactive Conversational Elicitation & Recommendation (ICER)
Moonkyung Ryu
Mohammad Ghavamzadeh
GENNEXT@SIGIR’25: The 1st Workshop on Next Generation of IR and Recommender Systems with Language Agents, Generative Models, and Conversational AI (2025)
Preview abstract
Large language models (LLM), despite their success in conducting natural conversations and solving various challenging NLP tasks, may not aim to understand and solicit a user’s preferences and suggest recommendations using the learned user preferences through the interactions. One primary challenge for pursuing this task is the lack of rich conversational recommendation data sets (of user/agent dialog conversations) that contain diverse preference elicitation scenarios on top of item recommendations. While several standard conversational recommender datasets do contain dialogs that the agents query users for preferences, they are often restricted to the use of particular key-phrases that limit users’ preference expressions w.r.t. particular item features (e.g., location, price, rating, etc). On the other hand, hiring human raters to create a personalized conversational recommender dataset that consists of rich preference queries (with the use of various abstract preference-describing attributes) and recommendations is a very expensive, error-prone, and time-consuming process because it requires data collection on numerous personalized recommendation scenarios, where each of them may involve exponential possibilities of preference elicitation interactions (with many different queries). We propose a synthetic data generation methodology that utilizes a synthetic recommendation dialog simulator that generates templatized dialogs and uses Gemini Ultra LLM to in-paint the templatized dialogs so that the dialogs become linguistically more natural. The simulator generates recommendation dialogs that align to the user preferences that are represented by embeddings.
View details
Project Euphonia: Advancing Inclusive Speech Recognition through Expanded Data Collection and Evaluation
Alicia Martín
Bob MacDonald
Katrin Tomanek
Frontiers in Language Sciences (2025)
Preview abstract
Speech recognition models, predominantly trained on standard speech, often exhibit lower accuracy for individuals with accents, dialects, or speech impairments. This disparity is particularly pronounced for economically or socially marginalized communities, including those with disabilities or diverse linguistic backgrounds. Project Euphonia, a Google initiative originally launched in English dedicated to improving Automatic Speech Recognition (ASR) of disordered speech, is expanding its data collection and evaluation efforts to include international languages like Spanish, Japanese, French and Hindi, in a continued effort to enhance inclusivity. This paper presents an overview of the extension of processes and methods used for English data collection to more languages and locales, progress on the collected data, and details about our model evaluation process, focusing on meaning preservation based on Generative AI.
View details
Origin-destination travel demand estimation: an approach that scales worldwide, and its application to five metropolitan highway networks
Christopher Bian
Yechen Li
Willa Ng
Bin Yan
Janny Zhang
2025
Preview abstract
Estimating Origin-Destination (OD) travel demand is vital for effective urban planning
and traffic management. Developing universally applicable OD estimation
methodologies is significantly challenged by the pervasive scarcity of high-fidelity traffic
data and the difficulty in obtaining city-specific prior OD estimates (or seed ODs), which
are often prerequisite for traditional approaches. Our proposed method directly
estimates OD travel demand by systematically leveraging aggregated, anonymized
statistics from Google Maps Traffic Trends, obviating the need for conventional census
or city-provided OD data. The OD demand is estimated by formulating a single-level,
one-dimensional, continuous nonlinear optimization problem with nonlinear equality
and bound constraints to replicate highway path travel times. The method achieves
efficiency and scalability by employing a differentiable analytical macroscopic network
model. This model by design is computationally lightweight, distinguished by its
parsimonious parameterization that requires minimal calibration effort and its capacity
for instantaneous evaluation. These attributes ensure the method's broad applicability
and practical utility across diverse cities globally. Using segment sensor counts from
Los Angeles and San Diego highway networks, we validate our proposed approach,
demonstrating a two-thirds to three-quarters improvement in the fit to segment count
data over a baseline. Beyond validation, we establish the method's scalability and
robust performance in replicating path travel times across diverse highway networks,
including Seattle, Orlando, Denver, Philadelphia, and Boston. In these expanded
evaluations, our method not only aligns with simulation-based benchmarks but also
achieves an average 13% improvement in it's ability to fit travel time data compared to
the baseline during afternoon peak hours.
View details
Matryoshka Model Learning for Improved Elastic Student Models
Chetan Verma
Cho-Jui Hsieh
Ngot Bui
Yang Zhang
Wen Chen
Xin Liu
Inderjit Dhillon
2025
Preview abstract
Industry-grade ML models are carefully designed to meet rapidly evolving serving constraints, which requires significant resources for model development. In this paper, we propose MatTA, a framework for training multiple accurate Student models using a novel Teacher-TA-Student recipe. TA models are larger versions of the Student models with higher capacity, and thus allow Student models to better relate to the Teacher model and also bring in more domain-specific expertise. Furthermore, multiple accurate Student models can be extracted from the TA model. Therefore, despite only one training run, our methodology provides multiple servable options to trade off accuracy for lower serving cost. We demonstrate the proposed method, MatTA, on proprietary datasets and models. Its practical efficacy is underscored by live A/B tests within a production ML system, demonstrating 20% improvement on a key metric. We also demonstrate our method on GPT-2 Medium, a public model, and achieve relative improvements of over 24% on SAT Math and over 10% on the LAMBADA benchmark.
View details
A personal health large language model for sleep and fitness coaching
Anastasiya Belyaeva
Zhun Yang
Nick Furlotte
Chace Lee
Erik Schenck
Yojan Patel
Jian Cui
Logan Schneider
Robby Bryant
Ryan Gomes
Allen Jiang
Roy Lee
Javier Perez
Jamie Rogers
Cathy Speed
Shyam Tailor
Megan Walker
Jeffrey Yu
Tim Althoff
Conor Heneghan
Mark Malhotra
Shwetak Patel
Shravya Shetty
Jiening Zhan
Daniel McDuff
Nature Medicine (2025)
Preview abstract
Although large language models (LLMs) show promise for clinical healthcare applications, their utility for personalized health monitoring using wearable device data remains underexplored. Here we introduce the Personal Health Large Language Model (PH-LLM), designed for applications in sleep and fitness. PH-LLM is a version of the Gemini LLM that was finetuned for text understanding and reasoning when applied to aggregated daily-resolution numerical sensor data. We created three benchmark datasets to assess multiple complementary aspects of sleep and fitness: expert domain knowledge, generation of personalized insights and recommendations and prediction of self-reported sleep quality from longitudinal data. PH-LLM achieved scores that exceeded a sample of human experts on multiple-choice examinations in sleep medicine (79% versus 76%) and fitness (88% versus 71%). In a comprehensive evaluation involving 857 real-world case studies, PH-LLM performed similarly to human experts for fitness-related tasks and improved over the base Gemini model in providing personalized sleep insights. Finally, PH-LLM effectively predicted self-reported sleep quality using a multimodal encoding of wearable sensor data, further demonstrating its ability to effectively contextualize wearable modalities. This work highlights the potential of LLMs to revolutionize personal health monitoring via tailored insights and predictions from wearable data and provides datasets, rubrics and benchmark performance to further accelerate personal health-related LLM research.
View details
Quartic Quantum Speedups for Planted Inference Problems
Alexander Schmidhuber
Ryan O'Donnell
Physical Review X, 15 (2025), pp. 021077
Preview abstract
We describe a quantum algorithm for the Planted Noisy kXOR problem (also known as sparse Learning Parity with Noise) that achieves a nearly quartic (4th power) speedup over the best known classical algorithm while also only using logarithmically many qubits. Our work generalizes and simplifies prior work of Hastings, by building on his quantum algorithm for the Tensor Principal Component Analysis (PCA) problem. We achieve our quantum speedup using a general framework based on the Kikuchi Method (recovering the quartic speedup for Tensor PCA), and we anticipate it will yield similar speedups for further planted inference problems. These speedups rely on the fact that planted inference problems naturally instantiate the Guided Sparse Hamiltonian problem. Since the Planted Noisy kXOR problem has been used as a component of certain cryptographic constructions, our work suggests that some of these are susceptible to super-quadratic quantum attacks.
View details
Preview abstract
Bandit Convex Optimization is a fundamental class of sequential decision-making problems, where the learner selects actions from a continuous domain and observes a loss (but not its gradient) at only one point per round. We study this problem in non-stationary environments, and aim to minimize the regret under three standard measures of non-stationarity: the number of switches S in the comparator sequence, the total variation Delta of the loss functions, and the path-length P of the comparator sequence. We propose a polynomial-time algorithm, Tilted Exponentially Weighted Average with Sleeping Experts (TEWA-SE), which adapts the sleeping experts framework from online convex optimization to the bandit setting. For strongly convex losses, we prove that TEWA-SE is minimax-optimal with respect to known S and Delta by establishing matching upper and lower bounds. By equipping TEWA-SE with the Bandit-over-Bandit framework, we extend our analysis to environments with unknown non-stationarity measures. For general convex losses, we introduce a second algorithm, clipped Exploration by Optimization (cExO), based on exponential weights over a discretized action space. While not polynomial-time computable, this method achieves minimax-optimal regret with respect to known S and Delta, and improves on the best existing bounds with respect to P.
View details
Not Like Us, Hunty: Measuring Perceptions and Behavioral Effects of Minoritized Anthropomorphic Cues in LLMs
Jeffrey Basoah
Daniel Chechelnitsky
Tao Long
Katharina Reinecke
Chrysoula Zerva
Kaitlyn Zhou
Maarten Sap
Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency, ACM (2025), pp. 710-745
Preview abstract
As large language models (LLMs) increasingly adapt and personalize to diverse sets of users, there is an increased risk of systems appropriating sociolects, i.e., language styles or dialects that are associated with specific minoritized lived experiences (e.g., African American
English, Queer slang). In this work, we examine whether sociolect usage by a LLM agent affects user reliance on its outputs and user perception (satisfaction, frustration, trust, and social presence). We designed and conducted user studies where 498 African American English (AAE) speakers and 487 Queer slang speakers performed a set of question-answering tasks with LLM-based suggestions in either standard American English (SAE) or their self-identified sociolect.
Our findings showed that sociolect usage by LLMs influenced both reliance and perceptions, though in some surprising ways. Results suggest that both AAE and Queer slang speakers relied more on the SAELM, and had more positive perceptions of the SAELM. Yet, only Queer slang speakers felt more social presence from the QSLM over the SAE one, whereas only AAE speakers preferred and trusted the SAELM over the AAE one. These findings emphasize the need to test for behavioral outcomes rather than simply assume that personalization would lead to a better and safer reliance outcome. They also highlight the nuanced dynamics of minoritized language in machine interactions, underscoring the need for LLMs to be carefully designed to respect cultural and linguistic boundaries while fostering genuine user engagement and trust.
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
Our Approach to Protecting AI Training Data
Cindy Muya
Jason Novak
Cindee Madison
Reiner Critides
Ben Kamber
Niha Vempati
Jeremy Wiesner
Google, Google, Google, 1600 Amphitheatre Parkway, Mountain View, CA, 94043 (2025) (2025)
Preview abstract
Google has over 25 years experience protecting data from inappropriate access and unauthorized use. In the era of AI, Google has extended these best practices in data protection to ensure that the right data is used the right way to train models. This paper presents a number of these best practices, describes how Google applies them in its systems, and describes how Google Cloud customers can use Google Cloud capabilities to implement these practices themselves.
Protecting data requires both technical controls to enable safe data use at scale, and governance processes to ensure that companies have visibility and control over how their data is used. This fundamentally requires: understanding data and ensuring it has sufficient metadata in the form of attributes, controlling the data and implementing policies to allow (or disallow) certain usage based on those attributes, transforming data to enable its usage in policy compliant ways, and human oversight and governance.
Protecting data in AI inherits these requirements and introduces new requirements to account for unique AI-specific risks including memorization/recitation and the costs of training foundational models. Meeting these new risks requires new capabilities including enhanced understanding of data and model lineage as well as an increased ability to control data usage through checks on data for policy compliance at the time a training job is configured before it is run.
This white paper offers an in-depth look at data protection best practices and Google’s data protection capabilities, and is one of a series of publications about Google's Secure AI Framework (SAIF). Building upon its secure development practices, Google has developed and deployed a number of capabilities to understand, control, and transform data in its infrastructure so that data is both protected and used appropriately. This involves robust annotation systems to represent metadata and enable granular understanding of data at both an item and dataset level, policy engines that evaluate machine readable policies on that data using the metadata attributes, and sensors to understand how data is flowing across Google’s systems and raise alerts when policy violations occur. Moreover, Google has developed de-identification and anonymization systems to transform data to make it policy compliant and safer to use for AI training.
View details
Preview abstract
Suppose Alice has a distribution ${P}$ and Bob has a distribution ${Q}$. Alice wants to draw a sample $a\sim {P}$ and Bob a sample $b \sim {Q}$ such that $a = b$ with as high of probability as possible. It is well-known that, by sampling from an optimal coupling between the distributions, Alice and Bob can achieve $\Pr[a = b] = 1 - D_{\text{tv}}({P},{Q})$, where $D_{\text{tv}}({P},{Q})$ is the total variation distance between ${P}$ and ${Q}$.
What if Alice and Bob must solve this same problem \emph{without communicating at all?} Perhaps surprisingly, with access to public randomness, they can still achieve $\Pr[a = b] \geq \frac{1 - D_{\text{tv}}({P},{Q})}{1 + D_{\text{tv}}({P},{Q})} \geq 1-2D_{\text{tv}}({P},{Q})$ using a simple protocol based on the Weighted MinHash algorithm. This bound was shown to be optimal in the worst-case by Bavarian, Ghazi, Haramaty, Kamath, Rivest, and Sudan [ToC 2020].
In this work, we revisit the ``communication-free coupling'' problem. We provide a simpler proof of the optimality result from [Bavarian et al., 2020]. Moreover we show that, while the \emph{worst-case} success probability of Weighted MinHash cannot be improved, an equally simple protocol based on Gumbel sampling offers a Pareto improvement: for every pair of distributions ${P}$ and ${Q}$, Gumbel sampling achieves an equal or higher value of $\Pr[a = b]$ than Weighted MinHash.
Importantly, this improvement translates to practice. We demonstrate an application of communication-free coupling to \emph{speculative decoding}, a recent method for accelerating autoregressive large language models [Leviathan, Kalman, Matias, ICML 2023].
We show that communication-free protocols can be used to contruct \emph{\CSD{}} schemes, which have the desirable property that their output is fixed given a fixed random seed, regardless of what drafter is used for speculation. In experiments on a language generation task, Gumbel sampling outperforms Weighted MinHash.
Code is available at \url{https://github.com/majid-daliri/DISD}.
Finally, we study the coupling problem in the setting where communication is \emph{bounded}, rather than completely eliminated. We describe a protocol that uses just $O(\log(n/\epsilon))$ bits of communication to achieve $\Pr[a = b] = 1 - D_{\text{tv}}({P},{Q}) - \epsilon$, i.e. to essentially match optimal coupling.
View details
