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 11426 publications
When Android Apps Request Location with a Reason: A Developer Study and User Behavior Experiment
May Smith
Aman Jain
Proceedings on Privacy Enhancing Technologies (2027)
Preview abstract
Recent reports have highlighted how mobile apps share user location data with third parties, risking user privacy and platform trust. Although location data is highly sensitive, when users grant apps location access, they may not know the full extent to which it is used. We study how requiring Android apps to show a reason for location access could impact developers, users, and the platform.
We surveyed 323 Android app developers and found most supported such a requirement. The majority said it would have a positive impact on user privacy, trust for apps, and trust for Android, where impact on user trust for Android correlated most strongly with support. Many developers also said the intervention would increase the number of users granting location access. Yet their open-ended comments also revealed consistent concerns, such as apps providing dishonest reasons and platform verification.
To study the impact on user behavior, we conducted a randomized controlled experiment with 2579 US Android users. We tested how users' decisions to grant location access were impacted by app type, whether reasons were included in the requests, and the content of the reasons, including monetization. We did not find the reasons impacted users' decisions; decisions were instead driven by app type and demographics. Yet we did find the reasons could have a positive impact on user perception for the platform when the reasons did not include using data for ads. Our findings provide insights into developers' willingness to implement privacy-enhancing changes, and expose limits to improving user privacy by simply adding information to user interfaces.
View details
Model Checking for Flexible Networking Protocols
Andrew Johnson
Sean Song
Proceedings of the 38th International Conference on Computer Aided Verification (CAV 2026)
Preview abstract
Operating a network is a daunting task. Operating one at a global scale, with stringent service objectives and requirements to be available during maintenance and failures, is even more so. At Google, we operate such a network. This paper details our experience applying formal methods to some of the networking protocols that are developed and maintained by in-house engineers. These protocols centrally route network traffic to respond to changes in demand, react to network failures, and allow for maintenance and upgrades.
We used formal methods to target a class of bugs stemming from unclear specifications, unintended system interactions, and logical errors at the specification level. We show how we modeled our protocols using an off-the-shelf model checker and a custom harness to scale the model horizontally. We were able to recreate several recent bugs and verify that the fixes implemented were correct. Finally, we present a method called state projection that we used to increase confidence in the coverage of our models, which we added to the TLC model checker for TLA+. We created 7 different TLA+ models and showed that they were effective at recreating bugs and verifying our fixes to those bugs.
View details
Multi-Agent Design: Optimizing Agents with Better Prompts and Topologies
Han Zhou
Shariq Iqbal
Ivan Vulić
Anna Korhonen
International Conference on Learning Representations (ICLR) (2026)
Preview abstract
Large language models (LLMs), employed as multiple agents that interact and collaborate with each other, have excelled at solving complex tasks. The agents are programmed with {prompts} that declare their functionality, along with the {workflows} that orchestrate interactions within a structured flow. Designing prompts and workflows for multi-agent systems is inherently complex, especially when addressing a new task. It often demands expert-level knowledge and involves significant trial and error. Gaining a deep understanding of the factors that contribute to effective multi-agent systems is essential for automating the entire process. Motivated by this, we first conduct an in-depth analysis of the design spaces for multi-agent systems, focusing on the impact of prompts, scaling the number of agents, and common types of agentic modules. Our findings reveal that top-performing systems often emerge from simpler design spaces, where prompts play a critical role in enhancing agent functionality and enabling more effective scaling. Based on the insights, we propose Multi-Agent System Search (MASS), a multi-stage optimization framework that performs the optimization in a pruned design space, with prompts and an influential subset of modules. We show that MASS-optimized multi-agent systems outperform existing alterntives by a substantial margin. Based on the MASS-found systems, we finally propose design principles behind building effective multi-agent systems.
View details
Peeking Ahead of the Field Study: Exploring VLM Personas as Support Tools for Embodied Studies in HCI
Xinyue Gui
Ding Xia
Mark Colley
Yuan Li
Vishal Chauhan
Anubhav Anubhav
Ehsan Javanmardi
Stela Hanbyeol Seo
Chia-Ming Chang
Manabu Tsukada
Takeo Igarashi
Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems (CHI 26)
Preview abstract
Field studies are irreplaceable but costly, time-consuming, and error-prone, which need careful preparation. Inspired by rapid-prototyping in manufacturing, we propose a fast, low-cost evaluation method using Vision-Language Model (VLM) personas to simulate outcomes comparable to field results. While LLMs show human-like reasoning and language capabilities, autonomous vehicle (AV)-pedestrian interaction requires spatial awareness, emotional empathy, and behavioral generation. This raises our research question: To what extent can VLM personas mimic human responses in field studies? We conducted parallel studies: 1) one real-world study with 20 participants, and 2) one video-study using 20 VLM personas, both on a street-crossing task. We compared their responses and interviewed five HCI researchers on potential applications. Results show that VLM personas mimic human response patterns (e.g., average crossing times of 5.25 s vs. 5.07 s) lack the behavioral variability and depth. They show promise for formative studies, field study preparation, and human data augmentation.
View details
Inference Perf: A Benchmarking Tool for GenAI Inference
Sachin Varghese
Jason Kramberger
Brendan Slabe
Chen Wang
Yuan Tang
Journal of Open Source Software (2026)
Preview abstract
Inference Perf is a generative AI (GenAI) inference performance benchmarking tool aimed at benchmarking and analyzing the performance of inference deployments. It is designed to be model-server agnostic, allowing for apples-to-apples comparisons across different model servers and serving stacks. As part of the inference benchmarking and metrics standardization effort in the Kubernetes wg-serving, it seeks to standardize tooling and metrics for measuring inference performance across the Kubernetes and model server communities.
View details
SpatialStack: Layered Geometry-Language Fusion for 3D VLM Spatial Reasoning
Jian Zhang
Bangya Liu
Achuta Kadambi
Zhiwen Fan
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2026)
Preview abstract
Large vision-language models (VLMs) still struggle with reliable 3D spatial reasoning, a core capability for embodied and physical AI systems. This limitation arises from their inability to capture fine-grained 3D geometry and spatial relationships. While recent efforts have introduced multi-view geometry transformers into VLMs, they typically fuse only the deep-layer features from vision and geometry encoders, discarding rich hierarchical signals and creating a fundamental bottleneck for spatial understanding. To overcome this, we propose SpatialStack, a general hierarchical fusion framework that progressively aligns vision, geometry, and language representations across the model hierarchy. Moving beyond conventional late-stage vision-geometry fusion, SpatialStack stacks and synchronizes multi-level geometric features with the language backbone, enabling the model to capture both local geometric precision and global contextual semantics. Building upon this framework, we develop VLM-SpatialStack, a model that achieves state-of-the-art performance on multiple 3D spatial reasoning benchmarks. Extensive experiments and ablations demonstrate that our multi-level fusion strategy consistently enhances 3D understanding and generalizes robustly across diverse spatial reasoning tasks, establishing SpatialStack as an effective and extensible design paradigm for vision-language-geometry integration in next-generation multimodal physical AI systems.
View details
LLM-Powered Analysis of IoT User Reviews: Tracking and Ranking Security and Privacy Concerns
Taufiq Islam Protick
Anupam Das
Proceedings of the International AAAI Conference on Web and Social Media (ICWSM) (2026)
Preview abstract
Being able to understand the security and privacy (S&P) concerns of IoT users brings benefits to both developers and users. To learn about users' views, we examine Amazon IoT reviews - one of the biggest IoT markets. This work presents a state-of-the-art methodology to identify and categorize reviews in which users express S&P concerns. We developed an automated pipeline by fine-tuning GPT-3.5-Turbo to build two models: the Classifier-Rationalizer-Categorizer and the Thematic Mapper. By leveraging dynamic few-shot prompting and the model's large context size, our pipeline achieved over 97% precision and recall, significantly outperforming keyword-based and classical ML methods. We applied our pipeline to 91K Amazon reviews about fitness trackers, smart speakers and cameras, over multiple years. We found that on average 5% contained S&P concerns, while security camera exhibited the highest prevalence at 10%. Our method detected significantly more S&P-relevant reviews than prior works: 15x more for fitness trackers, 29% more for smart speakers, and 70% more for cameras. Our longitudinal analysis reveals that concerns like surveillance and data control have persisted for years, suggesting limited industry progress. We demonstrate that across all device types, users consistently demand more precise control over what data is collected and shared. We uncover challenges in multi-user and multi-device interactions, identifying two previously unreported themes concerning inadequate controls for account separation and data access. These findings, ranging from broad persistent trends to specific instances of customer loss, offer actionable insights for developers to improve user satisfaction and trust.
View details
ARM MTE Performance in Practice
Taehyun Noh
Yingchen Wang
Tal Garfinkel
Mahesh Madhav
Mattan Erez
Shravan Narayan
Usenix Security (2026)
Preview abstract
Large Language Models utilizing reasoning techniques improve task performance but incur significant latency and token costs due to verbose generation. Existing automatic prompt optimization(APO) frameworks target task accuracy exclusively at the expense of generating long reasoning traces. We propose Cost-Regularized Optimization of Prompts (CROP), an APO method that introduces regularization on response length by generating textual feedback in addition to standard accuracy feedback. This forces the optimization process to produce prompts that elicit concise responses containing only critical information and reasoning. We evaluate our approach on complex reasoning datasets, specifically GSM8K, LogiQA and BIG-Bench Hard. We achieved an 80.6% reduction in token consumption while maintaining competitive accuracy, seeing only a nominal decline in performance. This presents a pragmatic solution for deploying token-efficient and cost-effective agentic AI systems in production pipelines.
View details
Spectral amplification for ground-state energy estimation of electronic structure in first quantization
Alicja Dutkiewicz
Alec White
Guang Hao Low
Albert Eugene DePrince III
Marika Kieferova
Dominic Berry
arXiv:2607.15358 (2026)
Preview abstract
We demonstrate an asymptotic gate complexity improvement in first-quantized ground-state energy estimation of electronic structure Hamiltonians in a plane wave basis by employing the sum-of-squares spectral gap amplification protocol. The improvement relies on identifying a sum-of-squares representation of the Hamiltonian which provides a lower bound certificate and low cost block encoding that leads to a provably lower quantum phase estimation gate cost. This is achieved by using a sum-of-squares operator generated by the total charge density operator resulting in a block encoding normalization improvement of $\lambda = \mathcal{O}\left(\eta\Delta^{-1.5}+\eta^{1.5}\Delta^{-1} \right)$ compared to prior work $\lambda = \mathcal{O}(\eta\Delta^{-2}+\eta^2\Delta^{-1})$ where $\eta$ is the number of electrons and $\Delta$ is the simulation grid spacing. The asymptotic reduction in block encoding normalization and similar block encoding costs to prior work is demonstrated to reduce resource estimates for materials and chemical systems by a factor of $2 - 44\times$ corresponding to the lowest cost estimates for \textit{ab initio} materials simulation.
View details
Preview abstract
We introduce AMS (Activation-based Model Scanner), a tool that detects modifications to safety training in language models by measuring the geometric structure of safety-relevant concepts in activation space. Safety training creates measurable separation between harmful and benign content classes; certain safety modifications collapse or rotate this structure, while others leave it intact. We validate AMS across 14 model configurations spanning 4 architecture families (Llama, Gemma, Qwen, Mistral) and four safety-modification categories (instruction-tuned, base, abliterated, uncensored fine-tunes). Leave-one-out cross-validation of thresholds achieves 71% accuracy (10/14); bootstrap 95% confidence intervals on σ point estimates have median width 3.4σ and a substantial fraction of cells cross the PASS threshold under resampling. We further measure behavioral compliance on 20 stratified JailbreakBench prompts per model and find that σ on the harmful-content concept predicts compliance with Pearson r=−0.546 ( p=0.043 ); the rank-order Spearman correlation is weaker ( ρ=−0.423 , p=0.13 ). The structural signal predicts behavior directionally but with meaningful noise. Mechanistic analysis identifies a four-class taxonomy of safety-training modifications distinguished by activation-space signature: 1) training removal collapses cluster separation (e.g., base models, Dolphin variants: 0.5– 1.4σ ); 2) weight-orthogonalization-style abliteration both collapses separation and rotates the refusal direction (Llama-3.1-abliterated: σ=3.33 , direction cos sim 0.30); 3) rotation-without-collapse abliteration preserves cluster separation while rotating the refusal direction (Gemma-2-9b-abliterated: σ=4.54 , direction cos sim 0.84); and 4) behavioral fine-tuning that preserves both magnitude and direction (DarkIdol-1.2-Uncensored: σ=5.45 , direction preserved, 97% behavioral compliance). 1) and 2) AMS’s Tier 1 σ -threshold detects classes; 3) Tier 2 direction-similarity verification detects class; and 4) Class is undetectable by activation-only probing and represents a documented failure mode of the approach. We discuss threshold calibration, limitations of single-run measurement, and the open problem of detecting behavioral-only safety modifications.
View details
Performance and User Response of Android's Smartphone-Based Alerts in the 2025 Marmara Ereğlisi Earthquake
Marc Stogaitis
Youngmin Cho
Richard Allen
Boone Spooner
Patrick Robertson
Greg Wimpey
Robert Bosch
Nivetha Thiruverahan
Steve Malkos
Alexei Barski
Tajinder Gadh
Nature Communications (2026)
Preview abstract
This study presents a comprehensive evaluation of Google’s Android Earthquake Alert (AEA) system during the Mw 6.2 Marmara Ereğlisi, Türkiye earthquake. AEA detected the event 5.31 seconds after its initiation, alerting over 16 million users. Warning times for weak shaking (MMI III) reached up to 150 seconds, with a median of 56 seconds. While near-source warning windows were shorter, the system achieved 90% true positives and 99% precision overall. The high density of the phone network enabled faster detection than traditional stations, even for this offshore epicenter. Feedback data shows AEA recipients were highly likely to take protective actions, such as "drop, cover, and hold on," or warn others. Timely alerts substantially increased user engagement, perceived usefulness, and future trust. These results highlight how crowd-sourced technology and behavioral insights can effectively enhance seismic resilience on a massive scale.
View details
ALF: Advertiser Large Foundation Model for Multi-Modal Advertiser Understanding
Sunny Rajagopalan
Alireza Golestaneh
Shubhra Chandra
Min Zhou
Jonathan Vronsky
Songbai Yan
2026
Preview abstract
We present ALF (Advertiser Large Foundation model), a multi-modal transformer architecture for understanding advertiser behavior and intent across text, image, video and structured data modalities. Through contrastive learning and multi-task optimization, ALF creates unified advertiser representations that capture both content and behavioral patterns. Our model achieves state-of-the-art performance on critical tasks including fraud detection, policy violation identification, and advertiser similarity matching. In production deployment, ALF reduces false positives by 90\% while maintaining 99.8\% precision on abuse detection tasks. The architecture's effectiveness stems from its novel combination of multi-modal transformations, intersample attention mechanism, spectrally normalized projections, and calibrated probabilistic outputs.
View details
Online Advertising with Spatial Interactions
Yifan Wang
Mingfei Zhao
Proceedings of the ACM Web Conference 2026, Association for Computing Machinery, New York, NY, USA, 213–224
Preview abstract
Online advertising platforms must decide how to allocate multiple ads across limited screen real estate, where each ad's effectiveness depends not only on its own placement but also on nearby ads competing for user attention. Such spatial externalities — arising from proximity, clutter, or crowding — can significantly alter welfare and revenue outcomes, yet existing auction and allocation models typically treat ad slots as independent or ordered along a single dimension. We introduce a new framework for spatial externalities in online advertising, in which the value of an ad depends on both its slot and the configuration of surrounding ads. We model ad slots as points in a metric space, and model an advertiser's value as a function of both their bid and a discount factor determined by the configuration of other displayed ads.
Within this framework, we analyze two natural models. For the Nearest-Neighbor model, where the value suppression depends only on the closest neighboring ad, we present a polynomial-time algorithm that achieves a constant approximation for the general case. We show that the allocation rule is monotone and can be implemented as a truthful mechanism. For a structured setting of 2D Euclidean space, we provide a PTAS. In contrast, for the Product-Distance model, where interference is aggregated multiplicatively across all neighbors, we establish a strong (and nearly-tight) hardness of approximation -- no polynomial-time algorithm can achieve any polynomial-factor approximation unless P=NP, via a reduction from Max-Independent-Set. Our results provide a foundation for reasoning about spatial externalities in ad allocation and for designing efficient, truthful mechanisms under such interactions.
View details
Preview abstract
Ideation, i. e., generating, capturing, organizing, and evaluating ideas, is foundational to creative practices; yet little is known about how blind and low vision ( BLV) creators engage in ideation. We conducted interviews with 20 BLV creators working in different domains including poetry, music production, tactile graphics, digital drawing, and mixed- media arts. Our findings reveal how participants established accessible creative workflows under social and institutional constraints, drew inspiration from diverse physical and digital sources, captured and organized ideas across text, audio, visual, and physical modalities, and evaluated their work through human and AI feedback. We discuss how ableist norms and accessibility breakdowns cascade across ideation pathways, progressively narrowing BLV individuals’ creative possibilities. Finally, we offer design considerations for building accessible ideation support tools that amplify BLV creators’ distinctive practices.
View details
