Background: African adolescents and young adults account for a disproportionate number of new HIV infections. There is an urgent need to identify scalable and cost-effective behavioral HIV prevention strategies for this population. Using a condom at first sex is associated with a higher likelihood of consistent use later. Tumaini (“Hope for the Future” in Swahili; Emory University) is a choose-your-own-adventure smartphone game that has been shown to reduce the risk of unprotected first sex by end line in a 45-month randomized controlled trial in western Kenya. Objective: This study aimed to assess the impact of Tumaini on proximal outcomes related to condom use at first sex (specifically, behavioral intentions, self-efficacy, attitudes, and knowledge) longitudinally across mid-adolescence in the above trial. Methods: Adolescent participants (n=996, mean baseline age 14, SD 0.56 years) were randomized 1:1 to receive either a smartphone loaded with Tumaini or an attention-control math game for 5 to 7 weeks at 3 time points (mean age 14.0, SD 0.56; 15.3, SD 0.55; and 16.0, SD 0.56 years, respectively). They completed a behavioral survey at 13 time points, through mean age 17.7 (SD 0.56) years. Using generalized estimating equations and controlling for age at baseline, we modeled mean scores (overall and stratified by gender) on a range of condom-related survey items over time to assess mean differences at specific time points. We applied appropriate Bonferroni corrections to inferences about cross-arm differences in mean changes relative to baseline at 4 time points (after each intervention period and at end line; α=.05/4) and within-arm mean changes relative to baseline at each of the 12 post-baseline time points (α=.05/12). Analyses were conducted as intent-to-treat. Results: At end line, 97.8% (n=974) of the sample had been retained. Participants in both arms dedicated a mean total of >30 hours to their assigned game. There was significant improvement across all condom-related proximal outcomes in the intervention arm relative to the control arm immediately after initial intervention exposure. For almost all outcomes, a significant cross-arm difference was also present at end line and for most outcomes at the 2 intervening comparison time points. Some outcomes saw stronger intervention effects on female participants (eg, self-efficacy to refuse unprotected sex) or male participants (eg, knowledge that condoms are an effective way to prevent HIV). In each arm, intention to use a condom at first sex was consistently higher among male participants; however, female intervention-arm scores overtook male control-arm scores following initial intervention exposure. Conclusions: Tumaini significantly improved theory-based proximal outcomes related to condom use, with effects sustained 45 months post initial exposure and 16 months post most recent exposure. Adolescents benefited from even short-term exposure, though repeated exposure generally sustained and reinforced intervention effects. As access to smartphones increases, Tumaini has potential for high scalability and impact on condom-related outcomes. Trial Registration: ClinicalTrials.gov NCT04437667; https://clinicaltrials.gov/study/NCT04437667

The multi-year deal involves at least a gigawatt of compute power and also includes a strategic investment from Nvidia.

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.

How AI is turning the Iran conflict into theater 

Much of the spotlight on AI in the Iran conflict has focused on models like Claude helping the US military decide where to strike. But a wave of “vibe-coded” intelligence dashboards—and the ecosystem surrounding them—reflect a new role that AI is playing in wartime: mediating information, often for the worse. 

These sorts of intelligence tools have much promise. Yet there are real reasons to be suspicious of their data feeds. Read the full story. 

—James O’Donnell 

This story is from The Algorithm, our weekly newsletter on AI. Sign up to receive it in your inbox every Monday. 

The must-reads 

I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology. 

1 Anthropic has sued the US government  
The AI firm wants to stop the Pentagon from blacklisting it. (Reuters) 
+ The White House is preparing a new executive order to weed out the company’s technology. (Axios) 
+ Defense experts are alarmed. (CNBC) 
+.Google and OpenAI staff have filed a legal brief backing Anthropic against Trump. (Wired $) 
+ The company’s stance won many supporters. (MIT Technology Review) 

2 GPS jamming has become a crucial battleground in the Middle East  
The interference is endangering—and protecting—ships and planes. (BBC) 
+ Signal jamming has made navigating the Strait of Hormuz even more difficult. (Bloomberg) 
+ Quantum navigation offers a potential solution. (MIT Technology Review)  

3 A tech journalist found his AI clone editing for Grammarly 
It’s providing AI-generated feedback “inspired by” real writers without their consent. (Platformer) 
+ Could ChatGPT do the jobs of journalists and copywriters? (MIT Technology Review) 

4 Nvidia plans to launch an open-source platform for AI agents  
It’s already pitching the “NemoClaw” product to enterprise software firms. (Wired $) 
+ But don’t let the AI agents hype get ahead of reality (MIT Technology Review) 
 
5 A startup wants to launch a space mirror that reflects sunlight onto Earth 
Reflect Orbital reckons it could power solar panels at night. Scientists are appalled. (NYT) 

6 Yann LeCun’s AI startup has raised over $1bn in Europe’s largest seed round  
Meta’s former chief AI scientist plans to build systems that “understand the world.” (Bloomberg) 

7 Hinge’s CEO insists the app doesn’t rate users’ attractiveness 
Jackie Jantos’ strategy has helped Hinge defy the decline in dating apps. (FT $) 
+ AI companions are stealing hearts—and it’s getting weird. (New Yorker $) 
+ It’s surprisingly easy to fall into a relationship with a chatbot. (MIT Technology Review) 

8 “AI psychosis” could be afflicting your loved ones  
If so, here’s how you can help them. (404 Media) 
+ One solution: AI should be able to “hang up” on you. (MIT Technology Review) 

9 Nintendo is suing Trump over illegal tariffs 
The gaming giant has joined a lawsuit seeking over $200 billion in refunds. (Ars Technica) 

10 Bio-tech is turning ancient poop into a map of lost civilizations  
Molecular sensors are finding human traces where physical ruins have vanished. (Nature)    

Quote of the day 

“I don’t think any of us, whether it’s me or Dario [Amodei], Sam Altman, or Elon Musk, has any legitimacy to decide for society what is a good or bad use of AI.”

—Yann LeCun gives Wired his take on the Anthropic’s spat the Pentagon. 

One More Thing 

This giant microwave may change the future of war 

armed forces are hunting for a weapon that disables drones en masse—and they want it fast.  

One solution focuses on microwaves: high-powered electronic devices that push out kilowatts of power to zap the circuits of a drone as if it were the tinfoil you forgot to take off your leftovers when you heated them up. 

Defense tech startup Epirus may have the winning formula. The company has developed a cutting-edge, cost-efficient drone zapper that’s sparking the interest of the US military. And drones are just one of its targets. Read the full story. 

—Sam Dean 

We can still have nice things 

A place for comfort, fun and distraction to brighten up your day. (Got any ideas? Drop me a line.) 

+ Werner Herzog’s magnificent movie about Africa’s ghost elephants has arrived on Disney+ and Hulu. 
+ A “city killer” asteroid won’t hit Earth after all. Phew.  
+ The Met is publishing high-definition 3D scans of over 100 iconic works. 
+ Marty and Doc from Back to the Future are still BFFs in real life. 

Top image credit: MIT TECHNOLOGY REVIEW (ILLUSTRATION) | PHOTO OF MISSILE (US NAVY), AI-GENERATED IMAGE OF RUBBLE VIA X, SCREENSHOTS VIA WORLDMONITOR, GLOBALTHREATMAP 

Send asteroids to hi@technologyreview.com.  

You can follow me on LinkedIn. Thanks for reading! 
 
 

—Thomas  

arXiv:2603.08322v1 Announce Type: new Abstract: We study mathematical discovery through the lens of neurosymbolic reasoning, where an AI agent powered by a large language model (LLM), coupled with symbolic computation tools, and human strategic direction, jointly produced a new result in combinatorial design theory. The main result of this human-AI collaboration is a tight lower bound on the imbalance of Latin squares for the notoriously difficult case $n \equiv 1 \pmod{3}$. We reconstruct the discovery process from detailed interaction logs spanning multiple sessions over several days and identify the distinct cognitive contributions of each component. The AI agent proved effective at uncovering hidden structure and generating hypotheses. The symbolic component consists of computer algebra, constraint solvers, and simulated annealing, which provides rigorous verification and exhaustive enumeration. Human steering supplied the critical research pivot that transformed a dead end into a productive inquiry. Our analysis reveals that multi-model deliberation among frontier LLMs proved reliable for criticism and error detection but unreliable for constructive claims. The resulting human-AI mathematical contribution, a tight lower bound of $4n(n{-}1)/9$, is achieved via a novel class of near-perfect permutations. The bound was formally verified in Lean 4. Our experiments show that neurosymbolic systems can indeed produce genuine discoveries in pure mathematics.

arXiv:2603.08388v1 Announce Type: new Abstract: We propose a Hierarchical Error-Corrective Graph FrameworkforAutonomousAgentswithLLM-BasedActionGeneration(HECG),whichincorporates three core innovations: (1) Multi-Dimensional Transferable Strategy (MDTS): by integrating task quality metrics (Q), confidence/cost metrics (C), reward metrics (R), and LLM-based semantic reasoning scores (LLM-Score), MDTS achieves multi-dimensional alignment between quantitative performance and semantic context, enabling more precise selection of high-quality candidate strate gies and effectively reducing the risk of negative transfer. (2) Error Matrix Classification (EMC): unlike simple confusion matrices or overall performance metrics, EMC provides structured attribution of task failures by categorizing errors into ten types, such as Strategy Errors (Strategy Whe) and Script Parsing Errors (Script-Parsing-Error), and decomposing them according to severity, typical actions, error descriptions, and recoverability. This allows precise analysis of the root causes of task failures, offering clear guidance for subsequent error correction and strategy optimization rather than relying solely on overall success rates or single performance metrics. (3) Causal-Context Graph Retrieval (CCGR): to enhance agent retrieval capabilities in dynamic task environments, we construct graphs from historical states, actions, and event sequences, where nodes store executed actions, next-step actions, execution states, transferable strategies, and other relevant information, and edges represent causal dependencies such as preconditions for transitions between nodes. CCGR identifies subgraphs most relevant to the current task context, effectively capturing structural relationships beyond vector similarity, allowing agents to fully leverage contextual information, accelerate strategy adaptation, and improve execution reliability in complex, multi-step tasks.

arXiv:2603.06585v1 Announce Type: cross Abstract: We present an open, sport-agnostic platform that turns tracking into comparable spatial measures across professional Ultimate, basketball, and soccer. Coaches in all three sports ask the same question: where is the usable space, and when should an off-ball run start? Our workflow standardizes inputs, provides timing-aware spatial evaluations, and makes it possible to reuse the same analysis across sports. We illustrate the approach with Ultimate as a focused testbed and then examine transfer between basketball and soccer. Together, these results show a practical path toward consistent, comparable evaluation across various invasion sports.

arXiv:2603.06590v1 Announce Type: cross Abstract: The Abstraction and Reasoning Corpus (ARC) is designed to assess generalization beyond pattern matching, requiring models to infer symbolic rules from very few examples. In this work, we present a transformer-based system that advances ARC performance by combining neural inference with structure-aware priors and online task adaptation. Our approach is built on four key ideas. First, we reformulate ARC reasoning as a sequence modeling problem using a compact task encoding with only 125 tokens, enabling efficient long-context processing with a modified LongT5 architecture. Second, we introduce a principled augmentation framework based on group symmetries, grid traversals, and automata perturbations, enforcing invariance to representation changes. Third, we apply test-time training (TTT) with lightweight LoRA adaptation, allowing the model to specialize to each unseen task by learning its transformation logic from demonstrations. Fourth, we design a symmetry-aware decoding and scoring pipeline that aggregates likelihoods across augmented task views, effectively performing ``multi-perspective reasoning'' over candidate solutions. We demonstrate that these components work synergistically: augmentations expand hypothesis space, TTT sharpens local reasoning, and symmetry-based scoring improves solution consistency. Our final system achieves a significant improvement over transformer baselines and surpasses prior neural ARC solvers, closing the gap toward human-level generalization.

arXiv:2603.06599v1 Announce Type: cross Abstract: Artificial intelligence pipelines -- spanning data collection, model training, deployment, and post-deployment monitoring -- concentrate ethical risks that intensify with multimodal and agentic systems. Existing governance instruments, including the EU AI Act, the IEEE 7000 series, and the NIST AI Risk Management Framework, provide high-level guidance but often lack enforceable, end-to-end operational controls. This paper presents an ethics-by-design control architecture that embeds consequentialist, deontological, and virtue-ethical reasoning into stage-specific enforcement mechanisms across the AI lifecycle. The framework implements a triple-gate structure at each lifecycle stage: Metric gates (quantitative performance and safety thresholds), Governance gates (legal, rights, and procedural compliance), and Eco gates (carbon and water budgets and sustainability constraints). It specifies measurable trigger conditions, escalation paths, audit artefacts, and mappings to EU AI Act obligations and NIST RMF functions, enabling integration with existing MLOps and CI/CD pipelines. Illustrative examples from large language model pipelines demonstrate how gate-based controls can surface and constrain technical, social, and environmental risks prior to release and during runtime. The framework is accompanied by a preregistered evaluation protocol that defines ex ante success criteria and assessment procedures, enabling falsifiable evaluation of gate effectiveness. By translating normative commitments into enforceable and testable controls, the framework provides a practical basis for operational AI governance across organizational contexts, jurisdictions, and deployment scales.

arXiv:2603.06600v1 Announce Type: cross Abstract: Vision Language Models (VLMs) are prone to errors, and identifying where these errors occur is critical for ensuring the reliability and safety of AI systems. In this paper, we propose an approach that automatically generates questions designed to deliberately induce incorrect responses from VLMs, thereby revealing their vulnerabilities. The core of this approach lies in fuzz testing and reinforcement finetuning: we transform a single input query into a large set of diverse variants through vision and language fuzzing. Based on the fuzzing outcomes, the question generator is further instructed by adversarial reinforcement fine-tuning to produce increasingly challenging queries that trigger model failures. With this approach, we can consistently drive down a target VLM's answer accuracy -- for example, the accuracy of Qwen2.5-VL-32B on our generated questions drops from 86.58\% to 65.53\% in four RL iterations. Moreover, a fuzzing policy trained against a single target VLM transfers to multiple other VLMs, producing challenging queries that degrade their performance as well.

arXiv:2603.06603v1 Announce Type: cross Abstract: Data duplication during pretraining can degrade generalization and lead to memorization, motivating aggressive deduplication pipelines. However, at web scale, it is unclear what constitutes a ``duplicate'': beyond surface-form matches, semantically equivalent documents (e.g. translations) may induce redundant training signals once models become sufficiently capable. Practically, this means that semantic duplicates operate increasingly like exact duplicates during training. We present evidence that duplication is scale-dependent in two ways. First, as model capability increases, cross-entropy loss gradients for semantically equivalent documents become more aligned. Smaller models, by contrast, produce gradients that reflect surface similarity (e.g., shared tokens) rather than semantic similarity. Second, we embedded all 192 million FineWeb-Edu-Dedup documents using EmbeddingGemma-300m. For moderate corpus sizes, the cosine similarity between nearest-neighbors follows an isotropic power law baseline. However, as corpus size grows to hundreds of billions of tokens, the nearest-neighbor similarities deviate sharply, indicating accelerated semantic collisions. Finally, controlled pretraining on data sampled with replacement from pools of finite unique documents shows that limited uniqueness yields mild degradation for small models, but rapidly increasing loss penalties for larger models, breaking naive scaling extrapolation. We derive explicit scaling laws that allow practitioners to estimate deviation from expected scaling due to limited semantic uniqueness of the pretraining corpus. Our results identify and resolve an unstudied source of scale-dependence, allowing for more accurate prediction at scale.

arXiv:2603.06612v1 Announce Type: cross Abstract: Pass@k and other methods of scaling inference compute can improve language model performance in domains with external verifiers, including mathematics and code, where incorrect candidates can be filtered reliably. This raises a natural question: can we similarly scale compute to elicit gains in truthfulness for domains without convenient verification? We show that across five benchmarks and models, surprisingly, it cannot. Even at 25x the inference cost of naive sampling, polling-style aggregation yields no consistent accuracy gains over single-sample baselines and often amplifies shared misconceptions. We find that under uncertainty, models are better at predicting what other models will say within model ensembles than at identifying what is true, revealing a separation between social prediction and truth verification. Across models and benchmarks, aggregation fails to provide a robust truth signal because language model errors are strongly correlated. The source of correlation goes beyond any individual benchmark: we show that even when conditioned on out of distribution random strings and asked to produce pseudo-random outputs, different models produce correlated outputs. Confidence-based weighting provides no benefit because self-reported confidence fails to reliably distinguish correct from incorrect answers. These results delineate a boundary for inference-time scaling: in verified domains, additional samples provide more candidates for a verifier to filter; in unverified domains, additional samples merely reinforce shared misconceptions.

arXiv:2603.06615v1 Announce Type: cross Abstract: For multivariate co-generation in scientific applications, we advocate pairwise block rather than joint modeling of all variables. This design mitigates the computational burden and data imbalance. To this end, we propose an Annealed Co-Generation (ACG) framework that replaces high-dimensional diffusion modeling with a low-dimensional diffusion model, which enables multivariate co-generation by composing pairwise variable generations. We first train an unconditional diffusion model over causal variables that are disentangled into pairs. At inference time, we recover the joint distribution by coupling these pairwise models through shared common variables, enabling coherent multivariate generation without any additional training. By employing a three-stage annealing process-Consensus, Heating, and Cooling-our method enforces consistency across shared common variables and progressively constrains each pairwise data distribution to lie on a learnable manifold, while maintaining high likelihood within each pair. We demonstrate the framework's flexibility and efficacy on two distinct scientific tasks: flow-field completion and antibody generation. All datasets and code will be made publicly available upon publication.

arXiv:2603.06649v1 Announce Type: cross Abstract: The coastal regions of the eastern and southern United States are impacted by severe storm events, leading to significant loss of life and properties. Accurately forecasting storm surge and wind impacts from hurricanes is essential for mitigating some of the impacts, e.g., timely preparation of evacuations and other countermeasures. Physical simulation models like the ADCIRC hydrodynamics model, which run on high-performance computing resources, are sophisticated tools that produce increasingly accurate forecasts as the resolution of the computational meshes improves. However, a major drawback of these models is the significant time required to generate results at very high resolutions, which may not meet the near real-time demands of emergency responders. The presented work introduces HURRI-GAN, a novel AI-driven approach that augments the results produced by physical simulation models using time series generative adversarial networks (TimeGAN) to compensate for systemic errors of the physical model, thus reducing the necessary mesh size and runtime without loss in forecasting accuracy. We present first results in extrapolating model bias corrections for the spatial regions beyond the positions of the water level gauge stations. The presented results show that our methodology can accurately generate bias corrections at target locations spatially beyond gauge stations locations. The model's performance, as indicated by low root mean squared error (RMSE) values, highlights its capability to generate accurate extrapolated data. Applying the corrections generated by HURRI-GAN on the ADCIRC modeled water levels resulted in improving the overall prediction on the majority of the testing gauge stations.

arXiv:2603.06663v1 Announce Type: cross Abstract: Recent advances in training-free visual prompting, such as Set-of-Mark, have emerged as a promising direction for enhancing the grounding capabilities of multimodal language models (MLMs). These techniques operate by partitioning the input image into object regions and annotating them with marks, predominantly boxes with numeric identifiers, before feeding the augmented image to the MLM. However, these approaches treat marked objects as isolated entities, failing to capture the relationships between them. On these premises, we propose Graph-of-Mark (GoM), the first pixel-level visual prompting technique that overlays scene graphs onto the input image for spatial reasoning tasks. We evaluate GoM across 3 open-source MLMs and 4 different datasets, conducting extensive ablations on drawn components and investigating the impact of auxiliary graph descriptions in the text prompt. Our results demonstrate that GoM consistently improves the zero-shot capability of MLMs in interpreting object positions and relative directions, improving base accuracy in visual question answering and localization up to 11 percentage points.

arXiv:2603.06710v1 Announce Type: cross Abstract: Mining specifications from execution traces presents an automated way of capturing characteristic system behaviors. However, existing approaches are largely restricted to Boolean abstractions of events, limiting their ability to express data-aware properties. In this paper, we extend mining procedures to operate over richer datatypes. We first establish candidate functions in our domain that cover the set of traces by leveraging Syntax Guided Synthesis (SyGuS) techniques. To capture these function applications temporally, we formalize the semantics of TSL$_f$, a finite-prefix interpretation of Temporal Stream Logic (TSL) that extends LTL$_f$ with support for first-order predicates and functional updates. This allows us to unify a corresponding procedure for learning the data transformations and temporal specifications of a system. We demonstrate our approach synthesizing reactive programs from mined specifications on the OpenAI-Gymnasium ToyText environments, finding that our method is more robust and orders of magnitude more sample-efficient than passive learning baselines on generalized problem instances.

arXiv:2603.06719v1 Announce Type: cross Abstract: The Dynamic Targeting (DT) mission concept is an approach to satellite observation in which a lookahead sensor gathers information about the upcoming environment and uses this information to intelligently plan observations. Previous work has shown that DT has the potential to increase the science return across applications. However, DT mission concepts must address challenges, such as the limited spatial extent of onboard lookahead data and instrument mobility, data throughput, and onboard computation constraints. In this work, we show how the performance of DT systems can be improved by using supplementary data streamed from geostationary satellites that provide lookahead information up to 35 minutes ahead of time rather than the 1 minute latency from an onboard lookahead sensor. While there is a greater volume of geostationary data, the search space for observation planning explodes exponentially with the size of the horizon. To address this, we introduce a hierarchical planning approach in which the geostationary data is used to plan a long-term observation blueprint in polynomial time, then the onboard lookahead data is leveraged to refine that plan over short-term horizons. We compare the performance of our approach to that of traditional DT planners relying on onboard lookahead data across four different problem instances: three cloud avoidance variations and a storm hunting scenario. We show that our hierarchical planner outperforms the traditional DT planners by up to 41% and examine the features of the scenarios that affect the performance of our approach. We demonstrate that incorporating geostationary satellite data is most effective for dynamic problem instances in which the targets of interest are sparsely distributed throughout the overflight.

arXiv:2603.06760v1 Announce Type: cross Abstract: Designing aerial robots for specialized tasks, from perching to payload delivery, requires tailoring their aerodynamic shape to specific mission requirements. For tasks involving wide flight envelopes, the usual sequential process of first determining the shape and then the motion planner is likely to be suboptimal due to the inherent nonlinear interactions between them. This limitation has been motivating co-design research, which involves jointly optimizing the aerodynamic shape and the motion planner. In this paper, we present a general-purpose, gradient-based, nested co-design framework where the motion planner solves an optimal control problem and the aerodynamic forces used in the dynamics model are determined by a neural surrogate model. This enables us to model complex subsonic flow conditions encountered in aerial robotics and to overcome the limited applicability of existing co-design methods. These limitations stem from the simplifying assumptions they require for computational tractability to either the planner or the aerodynamics. We validate our method on two complex dynamic tasks for fixed-wing gliders: perching and a short landing. Our optimized designs improve task performance compared to an evolutionary baseline in a fraction of the computation time.

arXiv:2603.06881v1 Announce Type: cross Abstract: Ferroelectric field-effect transistors (FeFET)-based vertical NAND (Fe-VNAND) has emerged as a promising candidate to overcome z-scaling limitations with lower programming voltages. However, the data retention of 3D Fe-VNAND is hindered by the complex interaction between charge detrapping and ferroelectric depolarization. Developing optimized device designs requires exploring an extensive parameter space, but the high computational cost of conventional Technology Computer-Aided Design (TCAD) tools makes such wide-scale optimization impractical. To overcome these simulation barriers, we present a Physics-Informed Neural Operator (PINO)-based AI surrogate model designed for high-efficiency prediction of threshold voltage (Vth) shifts and retention behavior. By embedding fundamental physical principles into the learning architecture, our PINO framework achieves a speedup exceeding 10000x compared to TCAD while maintaining physical accuracy. This study demonstrates the model's effectiveness on a single FeFET configuration, serving as a pathway toward modeling the retention loss mechanisms.

arXiv:2603.06976v1 Announce Type: cross Abstract: We present the first large-scale, cross-domain evaluation of document chunking strategies for dense retrieval, addressing a critical but underexplored aspect of retrieval-augmented systems. In our study, 36 segmentation methods spanning fixed-size, semantic, structure-aware, hierarchical, adaptive, and LLM-assisted approaches are benchmarked across six diverse knowledge domains using five different embedding models. Retrieval performance is assessed using graded relevance scores from a state-of-the-art LLM evaluator, with Normalised DCG@5 as the primary metric (complemented by Hit@5 and MRR). Our experiments show that content-aware chunking significantly improves retrieval effectiveness over naive fixed-length splitting. The top-performing strategy, Paragraph Group Chunking, achieved the highest overall accuracy (mean nDCG@5~0.459) and substantially better top-rank hit rates (Precision@1~24%, Hit@5~59%). In contrast, simple fixed-size character chunking as baselines performed poorly (nDCG@5

arXiv:2603.06987v1 Announce Type: cross Abstract: Deploying visuomotor robots at scale is challenging due to the potential for anomalous failures to degrade performance, cause damage, or endanger human life. Bimanual manipulators are no exception; these robots have vast state spaces comprised of high-dimensional images and proprioceptive signals. Explicitly defining failure modes within such state spaces is infeasible. In this work, we overcome these challenges by training a probabilistic, history informed, world model within the compressed latent space of a pretrained vision foundation model (NVIDIA's Cosmos Tokenizer). The model outputs uncertainty estimates alongside its predictions that serve as non-conformity scores within a conformal prediction framework. We use these scores to develop a runtime monitor, correlating periods of high uncertainty with anomalous failures. To test these methods, we use the simulated Push-T environment and the Bimanual Cable Manipulation dataset, the latter of which we introduce in this work. This new dataset features trajectories with multiple synchronized camera views, proprioceptive signals, and annotated failures from a challenging data center maintenance task. We benchmark our methods against baselines from the anomaly detection and out-of-distribution detection literature, and show that our approach considerably outperforms statistical techniques. Furthermore, we show that our approach requires approximately one twentieth of the trainable parameters as the next-best learning-based approach, yet outperforms it by 3.8% in terms of failure detection rate, paving the way toward safely deploying manipulator robots in real-world environments where reliability is non-negotiable.