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 11658 publications
A 3D Scene Graphs Survey: Open Challenges and Future Directions
Dennis Rotondi
Francesco Argenziano
Sebastian Koch
Nathan Hughes
Martin Büchner
Johanna Wald
Lukas Schmid
Daniele Nardi
Abhinav Valada
Liam Paul
Luca Carlone
Kai Arras
Annual Review of Control, Robotics, and Autonomous Systems (ARCRAS), 10 (2027) (to appear)
Preview abstract
3D Scene Graphs (3DSGs) have emerged as a powerful representation for spatial AI by combining geometric grounding with semantic and relational abstractions of the environment. Their expressiveness has made them relevant to a broad range of problems in robotics and computer vision, including mapping, task and motion planning, scene understanding, and many others. However, the field remains fragmented: different communities adopt distinct formulations, construction pipelines, and evaluation protocols, making it difficult to compare methods, identify common assumptions, and assess remaining challenges for robust real- world deployment. This survey provides a unified and critical review of 3DSGs, with particular emphasis on open challenges and future directions. We first formalize 3DSGs under a common definition and analyze the principal modeling choices that characterize existing formulations, including node and edge attributes, hierarchical structure, dynamic scene representations, and affordance-aware extensions. We then review how 3DSGs are constructed from raw sensory observations, covering both learning-oriented and construction-oriented systems. Finally, we examine downstream applications and evaluation strategies, from intrinsic graph quality to task-level performance. To support the community, we also provide a dedicated website that organizes and extends the surveyed works.
View details
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
Preview abstract
Modern transformers have been scaled up in size, data, training-time compute, and more recently test-time compute based on Chain-of-Thought (CoT), but the consequences of trading off these parameters are unknown. We construct explicit tradeoffs between two of these aspects: the number of layers in a transformer and the number of CoT tokens. Our first result is general: with mild assumptions and a single-token prompt, for any integer k, a transformer using soft attention and SwiGLU MLP layers with L layers and n CoT tokens can be converted to a transformer with at most ckL layers and ⌈n/k⌉ CoT tokens, and vice versa, although with an increase to the size of other parameters. We then a multi-token prompt path following task, finding that additional preprocessing from the multi-token prompt changes the depth-CoT relationship in this case. We construct an attention-only network with L layers and 2^(L−2) heads that can traverse 2^(L−1) − 1 edges along a path with a single CoT token, demonstrating at least an exponential relationship between increases in depth and reductions in the number of CoT tokens needed to traverse the entire path. This relationship for the path following task holds empirically as well, as demonstrated by training networks with up to 5 layers, with the tradeoff between depth and the maximum number of edge traversals in a single CoT token exactly matching our theoretical constructions.
View details
Autonomous Semantic Harmonization and Decoupled Staging for Enterprise Cloud CRM and ERP Data Migration
Subhani Shaik
Universal Library of Engineering and Technology (2026)
Preview abstract
Large-scale enterprise mergers, acquisitions (M&A), and cloud digital transformations routinely face severe operational deadlocks during the migration of historical, transactional, and customer data across heterogeneous cloud software ecosystems. Specifically, migrating deep relational object hierarchies from Customer Relationship Management (CRM) platforms (e.g., Salesforce) into financial transaction ledgers in Enterprise Resource Planning (ERP) systems (e.g., NetSuite, SAP, Oracle ERP) introduces cross-system impedance mismatch, relational orphaning, and data drift. Standard industry data migration tools—such as Salesforce Data Loader, NetSuite Bulk Upload, and point-to-point Extract, Transform, Load (ETL) scripts—rely on rigid, manual one-to-one field mappings and synchronous write pipelines that necessitate disruptive production cutover windows.
In this paper, we present the Autonomous Semantic Harmonization Framework (ASHF), a formal architectural protocol designed to govern high-stakes cloud CRM and ERP data migrations. ASHF replaces point-to-point scripting (O(N²) integration complexity) with a centralized Knowledge-Graph-Driven Semantic Resolution Hub (O(N) complexity) that autonomously maps conflicting relational entities across source and target schemas. Furthermore, ASHF introduces a Decoupled Quarantined Staging Engine coupled with an Asynchronous Data Triage and Total Rollback Verification Protocol, enabling continuous extraction and mathematical referential verification without interrupting live source systems. Empirical benchmarks across multi-tenant enterprise migrations demonstrate that ASHF achieves 100% relational data integrity (zero data loss), reduces human-in-the-loop data mapping design effort by 50%, cuts financial audit reconciliation times by 50%, and compresses end-to-end migration build schedules by 30% compared to standard linear ETL tools.
Index Terms— Cloud Data Migration, Enterprise Systems Unification, Customer Relationship Management (CRM), Enterprise Resource Planning (ERP), Knowledge Graphs, Zero-Downtime Migration, Semantic Harmonization.
View details
DDRop: Generic Memory Interposer Attacks on Confidential VMs by Dropping DDR5 Writes
Jesse Demeulemeester
Stefan Gloor
Patrick Jattke
David Oswald
Martin Thompson
Kaveh Razavi
Ingrid Verbauwhede
Jo Van Bulck
ACM Conference on Computer and Communications Security (CCS) (2026)
Preview abstract
Trusted Execution Environments (TEEs) are increasingly deployed in the cloud to protect sensitive workloads through hardwareenforced isolation, remote attestation, and transparent memory encryption. However, to meet memory performance and size demands, modern TEEs omit cryptographic freshness guarantees, leaving them vulnerable to replay attacks by adversaries with physical memory access. Prior work demonstrated low-cost active interposition attacks on DDR4 without requiring expensive specialized equipment, but these techniques do not extend to DDR5, where existing approaches are limited to passive ciphertext side-channel analysis and rely on bus downclocking to accommodate legacy memory bus analyzers.
We present the first low-cost (<200$) DDR5 RDIMM interposer capable of active fault injection at native speeds. By injecting targeted parity errors to silently discard cache line writebacks, we introduce DDRop, a new primitive that exploits the absence of cryptographic freshness to break the integrity of Intel TDX, Scalable SGX, and AMD SEV-SNP. Building on this primitive and targeting the APIs exposed by the TDX module and AMD Secure Processor, we show that adversaries can gain ciphertext access, copy arbitrary victim pages, and inject malicious secure page-table entries. We demonstrate end-to-end attacks on an up-to-date TDX platform, including forcing any TD into debug mode and forging attestation reports. While software-level mitigations, including timing-based interposer detection and API hardening, may reduce the attack surface, our results demonstrate that active DDR5 bus interposition is practical at low cost, highlighting the need for robust cryptographic memory integrity protections against physical adversaries.
View details
Preview abstract
Validating conversational artificial intelligence (AI) for regulated medical software applications may present challenges, as static test datasets and manual review may be limited in identifying emergent, conversational anomalies. A multi-agent AI system may be configured in a closed-loop for automated validation. The system can, for example, utilize an end user persona simulator agent to generate prompts for a target model and a domain /regulatory expert adjudicator agent to evaluate the target model’s responses against a configurable rubric. A meta-analysis agent can analyze anomalies to identify underlying vulnerabilities, which may then be used to programmatically synthesize new adversarial personas. This adaptive process can generate evidence to support regulatory compliance and continuous performance monitoring for medical software algorithms systems.
View details
Trust, but Validate the Instrument: Auditing AI-Generated RTL Verification Plans on Authored Security-Regression Proxies
Hang Xiao
Chuhong Xu
Mark Zhou
Iris Qian
Annabel Yi
2026
Preview abstract
AI-generated RTL verification plans can satisfy a
provider schema yet fail at the boundary to trusted execution.
We present SecTB-RTL, an auditable framework covering
31 tasks and 124 authored hardware-security regressions. A
deterministic non-AI baseline killed 36, 75, and 78 mutants at
increasing resource limits. C1-R2 then failed at provider schema
prevalidation. After an outcome-blind schema repair, a 1,860-
call run completed with 1,857 provider-accepted responses, but
only nine passed the production semantic validator because
the generation-facing and execution contracts disagreed. We
therefore preserve the run as an instrument-validation incident
and report no prompt-effect estimate. The incident directly shows
that provider/schema acceptance is not evidence of execution
validity; our contract likewise treats compilation and structural
coverage only as diagnostics until the exact persisted artifact
passes the production path. A subsequent follow-up is excluded
because it did not satisfy the preregistered evidence-completeness
gate and is treated only as future work. We release the benchmark,
failure-preserving contract, incident provenance, and governance
controls needed to prevent infrastructure behavior from being
misreported as model behavior.
View details
Rolling Shutter Relative Pose Estimation Made Practical
Daniel Barath
European Conference on Computer Vision (ECCV) (2026)
Preview abstract
Rolling shutter (RS) cameras equip virtually all consumer devices, yet RS-aware relative pose estimation has remained impractical: the state-of-the-art solver requires a minimum of 20 point correspondences, making RANSAC-based robust estimation prohibitively expensive due to the exponential dependence of the iteration count on the sample size. We make RS relative pose estimation practical by introducing affine correspondences (ACs) into the RS two-view geometry. We derive novel \emph{RS-corrected affine constraints} that account for the coupling between point perturbations and the row-dependent essential matrix, providing two equations per correspondence beyond the standard epipolar constraint. Building on these constraints, we develop a linearized algebraic solver that estimates pose and RS motion from only 7 ACs. The solver exploits the physical smallness of RS parameters to linearize the constraints, eliminates the 12 RS unknowns via null-space projection, and solves the remaining degree-20 system via action matrices in 1.2\,ms. On the TUM RS benchmark, our method achieves the best pose and RS parameter accuracy among all tested methods and, uniquely among RS solvers, provides accurate translational velocity estimates -- which are poorly conditioned from point correspondences alone due to a $\vec{v}$-$\vec{t}$ coupling. On the global-shutter EuRoC MAV dataset, the solver achieves comparable accuracy to the standard 5-point algorithm, demonstrating that it generalizes well to the GS setting. Code will be made public.
View details
Preview abstract
Advancements in large language models (LLMs) have made them increasingly useful for complex reasoning tasks which previously required domain experts. One such task is quality evaluation of query responses produced by a search engine. Evaluation generates metrics necessary to study the quality, impact, and usefulness of product changes and features. Typically, to compute evaluation metrics, human experts are asked to rate various attributes of search responses. This process is generally quite expensive and requires several days to complete. As an alternative, LLMs are now being used to perform rating tasks with lower costs and latency. In addition, many new metrics are being developed to evaluate Google's new AI-based offerings, which require ratings too. As a result, there is much higher demand for LLM rating prediction tasks in comparison with the allocated TPU (Tensor Processing Unit) budget. A larger portion of the company's TPU resources are reserved for serving live user traffic. In this paper, we present the AI Rater Service (AIRS), an inference pipeline that employs several software engineering techniques to generate AI ratings with high reliability and low latency. AIRS maximizes LLM inference throughput by optimizing TPU resource utilization across various evaluation workflows, while minimizing latency for higher priority tasks.
View details
Preview abstract
Large language models have achieved remarkable capabilities across domains, yet mechanisms underlying sophisticated reasoning continue to be explored1,2. Recent reasoning-reinforced models, including OpenAI’s o-series and DeepSeek-r1, outperform other merely instruction-tuned models on complex cognitive tasks3,4, attributed to extended test-time computation through longer chains of thought5. Here we show that enhanced reasoning emerges not from extended computation alone, but from the systematic simulation of complex, multi-agent interactions—a society of thought—which enables the deliberate diversification and debate among internal cognitive perspectives characterized by distinct personality traits and domain expertise. Through quantitative analysis using classified outputs and mechanistic interpretability methods applied to reasoning traces6–8, we find that reasoning models like DeepSeek-r1 exhibit much greater perspective diversity than baseline models, activating broader and more conflict between heterogeneous personality- and expertise-related features during reasoning. This multi-agent structure manifests in conversational behaviors including question-answering sequences, perspective shifts, and reconciliation of conflicting views, as well as in socio-emotional roles that characterize back-and-forth conversation, which together account for over 60% of the accuracy advantage in reasoning tasks through both direct and indirect facilitation of cognitive strategies9,10. Controlled reinforcement learning experiments further reveal that priming models with conversational scaffolding—even when dialogues lead to incorrect solutions—substantially accelerates reasoning improvement compared to answer-only training. These findings indicate that the social organization of thought, rather than correctness alone, enables effective exploration of solution spaces. We suggest that reasoning models establish a computational parallel to collective intelligence in human groups11–13, where diversity enables superior problem-solving when systematically structured and suggest new opportunities for agent organization to harness the wisdom of crowds.
View details
A Dichotomy for Cubic Bipartite Holant Problems with Complex Algebraic Weights
arXiv preprint arXiv:2609.33473 (2026)
Preview abstract
We classify the exact evaluation of Holant(f | =_3) for every fixed complex algebraic symmetric Boolean ternary signature f. An input is a cubic bipartite multigraph: every vertex on one side carries f, every vertex on the other side carries ternary equality, and no auxiliary signatures are freely available. The tractable signatures are precisely rank-one tensors, generalized equalities, and equality-preserving cube-root diagonal transformations of six affine signatures, together with nonzero scalings and reversal. Every other signature gives a #P-hard problem under polynomial-time Turing reductions. We also identify the exact real intersection: it consists of the same tractable families as in the rational classification, with real algebraic parameters. The proof preserves degree exactly three on both sides of every oracle instance. A rank-one matrix extracted by interpolation supplies one unary signature only after its unused factor has been absorbed in triples. Over the complex numbers this absorption has three exceptional projective directions. We combine this constraint with projective matrix-group orbits, explicit ternary replacements, and an exhaustive treatment of finite projective orders. The cases of orders three and five include exact polynomial certificates; the certificate identities and a rational-arithmetic verifier are supplied as supplementary material.
View details
Preview abstract
The management of a hybrid workforce comprising human and autonomous computational agents may be challenged by the use of separate systems for human capital and software assets, which can create a governance gap. A system can provide a unified framework for managing a hybrid workforce. For example, the system may utilize a labor service mesh to analyze and route tasks to either a human intent tier or an agentic execution tier. A potential principle of the system is structural symmetry, where computational agents can be assigned digital identities and managed through a lifecycle process that may parallel human resource functions, such as onboarding, performance evaluation, and structured offboarding. This integrated approach can facilitate a unified system of record and governance model for an organization's intelligence capacity.
View details
Preview abstract
This paper demonstrates that artificial intelligence can accelerate mathematical discovery by autonomously solving an open problem in theoretical physics. We present a neuro-symbolic system, combining the Gemini Deep Think large language model with a systematic Tree Search (TS) framework and automated numerical feedback, that successfully derived novel, exact analytical solutions for the power spectrum of gravitational radiation emitted by cosmic strings. Specifically, the agent evaluated the core integral for arbitrary loop geometries, directly improving upon recent AI-assisted attempts that only yielded partial asymptotic solutions. To substantiate our methodological claims regarding AI-accelerated discovery and to ensure transparency, we detail system prompts, search constraints, and intermittent feedback loops that guided the model. The agent identified a suite of 6 different analytical methods, the most elegant of which expands the kernel in Gegenbauer polynomials to naturally absorb the integrand's singularities. The methods lead to an asymptotic result for at large that both agrees with numerical results and also connects to the continuous Feynman parameterization of Quantum Field Theory. We detail both the algorithmic methodology that enabled this discovery and the resulting mathematical derivations.
View details
Estimating high-resolution albedo for urban applications
Elizabeth Wesley
Salil Banerjee
Aniruddh Chennapragada
Kevin Crossan
Bryce Cronkite-Ratcliff
Ellie Delich
Tristan Goulden
Mansi Kansal
Eric Mackres
Yael Mayer
Rebecca Milman
Shravya Shetty
Charlotte Stanton
Wayne Sun
Lucy Hutyra
Nature Communications (2026)
Preview abstract
Extreme urban heat poses a growing threat to global health, yet a critical cooling intervention-reflective cool roofs-lacks the high-resolution data needed for building-level planning. While satellite-based albedo estimates are globally available, their coarse resolution cannot resolve individual rooftops, limiting targeted climate action. Here we show that fusing freely available 10-m Sentinel-2 imagery with high-resolution satellite data enables urban albedo mapping at 30-cm resolution. This approach combines the radiometric accuracy and global coverage of Sentinel-2 with the spatial detail of commercial imagery to produce actionable urban datasets. Validated against airborne hyperspectral measurements (root mean square error = 0.04), the method resolves albedo at individual building footprints across diverse urban environments. Applied to 12 global cities, our analysis shows that prioritizing large-footprint buildings for cool roof retrofits can reduce citywide temperatures by up to 0.5 °C. This work validates Sentinel-2 for city-scale albedo estimation and enables globally deployable, building-scale rooftop albedo mapping to support targeted cool roof interventions.
View details
WAXAL: A large-scale multilingual African language speech corpus
Abdoulaye Diack
Perry Nelson
MohamedElfatih MohamedKhair
Emmanuel Asiedu Brempong
Tavonga Siyavora
Bob MacDonald
Uche Okonkwo
Sandy Ritchie
Mandy Jordan
Abhishek Bapna
Vusumuzi Dube
Jason Hickey
Ronit Levavi Morad
Jeff Dean
Aisha Walcott-Bryant
2026
Preview abstract
Recent advances in speech technology predominantly favor high-resource languages, leaving speakers of Sub-Saharan African languages underserved. To bridge this divide, we introduce WAXAL, an open, large-scale speech dataset spanning 24 African languages representing over 100 million speakers. WAXAL comprises over 14,100 hours of speech collected in partnership with African academic and community organizations to build local research capacity: an Automatic Speech Recognition (ASR) corpus with 2,340 hours of transcribed, image-prompted natural speech across 19 languages (plus ~11,760 untranscribed hours) and a studio-quality Text-to-Speech (TTS) corpus exceeding 235 hours. We empirically benchmark three architecturally distinct ASR models---Gemma 3n 2B, Whisper-Large-v3, and MMS-1B-all---across the 19 ASR languages on strictly speaker-disjoint splits. Fine-tuning on WAXAL yields substantial gains, reducing macro-average WER by up to 58% for Whisper-Large-v3, 56% for Gemma 3n 2B, and 17% for MMS-1B-all (and Character Error Rate by up to 70%). WAXAL-fine-tuned models generalize zero-shot to three out-of-distribution benchmarks---FLEURS, Mozilla Common Voice, and Sunbird SALT---and outperform FLEURS-trained baselines in bidirectional cross-dataset transfer. Finally, an LLM-as-judge semantic evaluation confirms that fine-tuning preserves speaker intent beyond surface-level transcription metrics. WAXAL is publicly released under a CC-BY-4.0 license at https://huggingface.co/datasets/google/WaxalNLP
View details
