Diffusion-based visuomotor policies deployed with asynchronous inference often exhibit inter-chunk discontinuities and lack explicit mechanisms for obstacle-aware execution, leading to jerky motions and collisions that hinder reliable manipulation in real-world scenes. To address these issues, we propose LAGO Policy, a unified asynchronous action-generation framework that integrates trajectory optimization with diffusion policy for smooth and safe execution. LAGO Policy improves inter-chunk consistency via latency-aware classifier-free guidance conditioning on future actions. It further enables goal-directed collision-free trajectory planning by predicting a task-relevant interaction goal from demonstrations. Finally, spatial-temporal trajectory optimization refines the actions to be executed for low-jerk and feasible motion. Extensive real-world experiments demonstrate that LAGO Policy achieves smooth collision-free execution with high task success across challenging manipulation tasks. Project Website: https://lago-policy.github.io/
The Learn-As-you-Go (LAGO) design is an adaptive clinical trial design that allows modifications to multicomponent intervention packages across stages. Centers participate in more than one stage, as is common in large-scale implementation trials. In LAGO trials, center characteristics may act as confounders, predicting both the intervention package and the outcomes. We extend the LAGO theory by introducing fixed center effects to control for confounding by indication through measured and unmeasured center characteristics. Conditioning on center characteristics by including fixed center effects ensures asymptotic results hold without requiring explicit characterization of unmeasured confounders. Our methods apply even with small numbers of centers. LAGO theory is established for continuous outcomes following a generalized linear model and binary outcomes following a logistic regression model, unifying theory across outcome types. Point- and interval estimators are derived, and consistency and asymptotic normality are established. Valid hypothesis tests for the overall intervention effect are provided, and the optimal intervention package minimizing cost subject to a target outcome m
Large language models (LLMs) have shown strong potential for planning and sequential decision-making, but prior work often relies on using them as direct controllers, which requires precise action generation and can be unreliable in practice. This paper proposes Latent Action Guidance for Online Reinforcement Learning (LaGO), a framework that uses a pretrained LLM as a latent action prior to softly guide online policy optimization, rather than treating the LLM as an explicit planner or controller. Experiments on both a discrete-control benchmark, CLEVR-Robot, and a continuous-control benchmark, Meta-World, demonstrate that LaGO consistently improves both reward and success rate over Vanilla PPO. In particular, LaGO increases the average success rate from 15.1% to 27.2% on CLEVR-Robot and from 2.7% to 15.2% on Meta-World. Our analysis further shows that stronger pretrained LLMs provide more effective guidance, suggesting that LLM knowledge can improve planning and online decision-making.
We introduce LAGO, a LocAl-Global Optimization framework coupling Bayesian Optimization (BO) and gradient-based trust region local refinement through an adaptive competition mechanism for smooth expensive-to-evaluate objective functions with available gradients. At each iteration, global and local optimization strategies independently propose candidate points, and the next evaluation is selected based on predicted improvement. LAGO separates global exploration from local refinement at the proposal level: the BO acquisition function is optimized outside the active trust region, while local candidates are proposed within the trust region. Points in the vicinity of the accepted local step are incorporated in the global GP dataset only when satisfying a lengthscale-based minimum-distance criterion, hence reducing the risk of numerical instability during local exploitation. LAGO enhances BO with efficient local refinement when reaching promising regions, and reverts to exploratory behavior when local steps are not competitive.
In the face of vast numbers of preventable deaths worldwide and gaping disparities in their distribution, we cannot afford to conduct null and inconclusive effectiveness and implementation trials of evidence-based interventions. The gold standard in biomedical research, the individually randomized clinical trial, is ill-suited as the primary tool for knowledge generation for contextually relevant, scalable, complex public health interventions of multi-component strategies. In this paper, we discuss the new Learn-As-you-GO (LAGO) design. In LAGO trials, the components of a complex intervention package are repeatedly optimized in pre-planned stages, until the package achieves its outcome and power goals with minimized cost and/or other optimization criteria, such as maximizing patient satisfaction. In this paper, the inputs to, and outputs of, LAGO are described, along with its general methodology. The methods are illustrated in the BetterBirth study, a large trial that aimed to reduce maternal and neonatal mortality in Uttar Pradesh, India, using the WHO essential birth practices checklist. Despite its scale, the BetterBirth study failed to demonstrate a significant effect of the in
The Learn-As-you-GO (LAGO) design provides a rigorous framework for adapting the intervention package based on accumulating data while the trial is ongoing. This article improves the flexibility of the LAGO design by incorporating statistical power as an optimization criterion (power goal) in LAGO optimizations. We propose the unconditional and conditional power approaches to add a power goal. Both approaches estimate the power at the end of the LAGO trial using data from prior stages, and increase the power at the end of the LAGO trial when the original trial was underpowered. Including a power goal maintains the asymptotic properties of the estimators of the treatment effect while preserving the asymptotic level of the statistical test at the end of the trial. We illustrate the benefits of our methods through a retrospective application to the BetterBirth Study, a large-scale study of maternal-newborn care that failed to show a significant effect on its primary outcome. This analysis demonstrates how our methods could have led to more intensive interventions and potentially significant results. The LAGO design with power goal optimizations provides investigators with a powerful t
We propose LAGO - Language Similarity-Aware Graph Optimization - a novel approach for few-shot cross-lingual embedding inversion attacks, addressing critical privacy vulnerabilities in multilingual NLP systems. Unlike prior work in embedding inversion attacks that treat languages independently, LAGO explicitly models linguistic relationships through a graph-based constrained distributed optimization framework. By integrating syntactic and lexical similarity as edge constraints, our method enables collaborative parameter learning across related languages. Theoretically, we show this formulation generalizes prior approaches, such as ALGEN, which emerges as a special case when similarity constraints are relaxed. Our framework uniquely combines Frobenius-norm regularization with linear inequality or total variation constraints, ensuring robust alignment of cross-lingual embedding spaces even with extremely limited data (as few as 10 samples per language). Extensive experiments across multiple languages and embedding models demonstrate that LAGO substantially improves the transferability of attacks with 10-20% increase in Rouge-L score over baselines. This work establishes language simi
Zero-shot recognition aims to classify an image by selecting the most compatible label description from a set of candidate classes without any task-specific supervision. In fine-grained settings, however, the relevant evidence often lies in localized parts, attributes, or textures rather than in the full image, making whole-image alignment suboptimal. Recent localized visual-text alignment methods address this by comparing class descriptions with multiple image regions, but they typically rely on large sets of random or redundant crops, increasing inference cost and introducing many highly redundant or weakly relevant candidates. Moreover, introducing semantic guidance too early can create an error-amplifying feedback process in which inaccurate intermediate predictions bias later localization and reinforce subsequent mistakes; we refer to this failure mode as the prediction loop. We propose LAGO (LAnguage-Guided adaptive Object-region focus), a framework for efficient and robust zero-shot localized visual-text alignment. LAGO first performs class-agnostic object-centric candidate discovery to obtain a stable visual initialization, and then applies adaptive language-guided refineme
LAGO, the Latin American Giant Observatory, is an extended cosmic ray observatory, consisting of a wide network of water Cherenkov detectors located in 10 countries. With different altitudes and geomagnetic rigidity cutoffs, their geographic distribution, combined with the new electronics for control, atmospheric sensing and data acquisition, allows the realisation of diverse astrophysics studies at a regional scale. It is an observatory designed, built and operated by the LAGO Collaboration, a non-centralised alliance of 30 institutions from 11 countries. While LAGO has access to different computational frameworks, it lacks standardised computational mechanisms to fully grasp its cooperative approach. The European Commission is fostering initiatives aligned to LAGO objectives, especially to enable Open Science and its long-term sustainability. This work introduces the adaptation of LAGO to this paradigm within the EOSC-Synergy project, focusing on the simulations of the expected astrophysical signatures at detectors deployed at the LAGO sites around the World.
It is well known that changing the intervention package while a trial is ongoing does not lead to valid inference using standard statistical methods. However, it is often necessary to adapt, tailor, or tweak a complex intervention package in public health implementation trials, especially when the intervention package does not have the desired effect. This article presents conditions under which the resulting analyses remain valid even when the intervention package is adapted while a trial is ongoing. Our results on such Learn-As-you-GO (LAGO) studies extend the theory of LAGO for binary outcomes following a logistic regression model (Nevo, Lok and Spiegelman, 2021) to LAGO for continuous outcomes under flexible conditional mean model. We derive point and interval estimators of the intervention effects and ensure the validity of hypothesis tests for an overall intervention effect. We develop a confidence set for the optimal intervention package, which achieves a pre-specified mean outcome while minimizing cost, and confidence bands for the mean outcome under all intervention package compositions. This work will be useful for the design and analysis of large-scale intervention trial
In this paper, we present a method to automatically build large labeled datasets for the author ambiguity problem in the academic world by leveraging the authoritative academic resources, ORCID and DOI. Using the method, we built LAGOS-AND, two large, gold-standard datasets for author name disambiguation (AND), of which LAGOS-AND-BLOCK is created for clustering-based AND research and LAGOS-AND-PAIRWISE is created for classification-based AND research. Our LAGOS-AND datasets are substantially different from the existing ones. The initial versions of the datasets (v1.0, released in February 2021) include 7.5M citations authored by 798K unique authors (LAGOS-AND-BLOCK) and close to 1M instances (LAGOS-AND-PAIRWISE). And both datasets show close similarities to the whole Microsoft Academic Graph (MAG) across validations of six facets. In building the datasets, we reveal the variation degrees of last names in three literature databases, PubMed, MAG, and Semantic Scholar, by comparing author names hosted to the authors' official last names shown on the ORCID pages. Furthermore, we evaluate several baseline disambiguation methods as well as the MAG's author IDs system on our datasets, and
The Latin American Giant Observatory (LAGO) is an extended cosmic ray observatory composed of a network of water-Cherenkov detectors (WCD) spanning over different sites located at significantly different altitudes (from sea level up to more than $5000$\,m a.s.l.) and latitudes across Latin America, covering a wide range of geomagnetic rigidity cut-offs and atmospheric absorption/reaction levels. The LAGO WCD is simple and robust, and incorporates several integrated devices to allow time synchronization, autonomous operation, on board data analysis, as well as remote control and automated data transfer. This detection network is designed to make detailed measurements of the temporal evolution of the radiation flux coming from outer space at ground level. LAGO is mainly oriented to perform basic research in three areas: high energy phenomena, space weather and atmospheric radiation at ground level. It is an observatory designed, built and operated by the LAGO Collaboration, a non-centralized collaborative union of more than 30 institutions from ten countries. In this paper we describe the scientific and academic goals of the LAGO project - illustrating its present status with some re
We present the LAGOVirtual Project: an ongoing project to develop platform to collaborate in the Large Aperture GRB Observatory (LAGO). This continental-wide observatory is devised to detect high energy (around 100 GeV) component of Gamma Ray Bursts, by using the single particle technique in arrays of Water Cherenkov Detectors (WCD) at high mountain sites (Chacaltaya, Bolivia, 5300 m a.s.l., Pico Espejo, Venezuela, 4750 m a.s.l., Sierra Negra, Mexico, 4650 m a.s.l). This platform will allow LAGO collaboration to share data, and computer resources through its different sites. This environment has the possibility to generate synthetic data by simulating the showers through AIRES application and to store/preserve distributed data files collected by the WCD at the LAGO sites. The present article concerns the implementation of a prototype of LAGO-DR adapting DSpace, with a hierarchical structure (i.e. country, institution, followed by collections that contain the metadata and data files), for the captured/simulated data. This structure was generated by using the community, sub-community, collection, item model; available at the DSpace software. Each member institution-country of the pro
To characterize the signals registered by the different types of water Cherenkov detectors (WCD) used by the Latin American Giant Observatory (LAGO) Project, it is necessary to develop detailed simulations of the detector response to the flux of secondary particles at the detector level. These particles are originated during the interaction of cosmic rays with the atmosphere. In this context, the LAGO project aims to study the high energy component of gamma rays bursts (GRBs) and space weather phenomena by looking for the solar modulation of galactic cosmic rays (GCRs). Focus in this, a complete and complex chain of simulations is being developed that account for geomagnetic effects, atmospheric reaction and detector response at each LAGO site. In this work we shown the first steps of a GEANT4 based simulation for the LAGO WCD, with emphasis on the induced effects of the detector internal diffusive coating.
The Latin American Giant Observatory (LAGO) is a distributed cosmic ray observatory at a regional scale in Latin America, by deploying a large network of Water Cherenkov detectors (WCD) and other astroparticle detectors in a wide range of latitudes from Antarctica to México, and altitudes from sea level to more than 5500 m a.s.l. Detectors telemetry, atmospherics conditions and flux of secondary particles at the ground are measured with extreme detail at each LAGO site by using our own-designed hardware and firmware (ACQUA). To combine and analyse all these data, LAGO developed ANNA, our data analysis framework. Additionally, ARTI, a complete framework of simulations designed to simulate the expected signals at our detectors coming from primary cosmic rays entering the Earth atmosphere, allowing a precise characterization of the sites in realistic atmospheric, geomagnetic and detector conditions. As the measured and synthetic data started to flow, we are facing challenging scenarios given a large amount of data emerging, performed on a diversity of detectors and computing architectures and e-infrastructures. These data need to be transferred, analyzed, catalogued, preserved, and pr
The Latin American Giant Observatory (LAGO) is an extended astroparticle observatory with the goal of studying Gamma Ray Bursts (among other extreme universe phenomena), space weather and atmospheric radiation at ground level. It consists of a network of several Water Cherenkov Detectors (WCD) located at different sites and different latitudes along the American Continent (from Mexico up to the Antarctic region). Another interest of LAGO is to encourage and support the development of experimental basic research in Latin America, mainly with low cost equipment. In the case of Chiapas, Mexico, the experimental astroparticle physics activity was limited, up to now, to data analysis from other detectors located far away from the region. Thanks to the collaboration within LAGO, the deployment of one WCD is ongoing at the Universidad Autónoma de Chiapas (UNACH). This will allow, for the first time in the region, to train students and researchers in the deployment processes. Till now the setup of the signal-processing electronics has been performed and the characterization of the photomultiplier tube is currently being done. The main, short-term goal is to install one WCD on top of the Ta
We introduce a novel intersection type system for a $λ$-calculus with algebraic effects and handlers. The system, inherently behavioral in nature, enjoys the classical properties of intersection type systems, in particular subject reduction and expansion. It thus characterizes the set of terms whose evaluation process terminates and, at the same time, allows reducing the reachability problem to type inference. This new system, the first with these features for a calculus with handlers, induces a system of simple types which, although not guaranteeing termination, is type sound and admits a decidable HOMC problem, unlike similar type systems like Dal Lago and Ghyselen's HEPCF.
Mechanistic interpretability has transformed the analysis of transformer circuits by decomposing model behavior into competing algorithms, identifying phase transitions during training, and deriving closed-form predictions for when and why strategies shift. However, this program has remained largely confined to sequence-prediction architectures, leaving embodied control systems without comparable mechanistic accounts. Here we extend this framework to sensorimotor-cognitive development, using infant motor learning as a model system. We show that foundational inductive biases give rise to causal control circuits, with learned gating mechanisms converging toward theoretically motivated uncertainty thresholds. The resulting dynamics reveal a clean phase transition in the arbitration gate whose commitment behavior is well described by a closed-form exponential moving-average surrogate. We identify context window k as the critical parameter governing circuit formation: below a minimum threshold (k$\leq$4) the arbitration mechanism cannot form; above it (k$\geq$8), gate confidence scales asymptotically as log k. A two-dimensional phase diagram further reveals task-demand-dependent route a
With growing discussions about the carbon footprint of academic conferences, more questions are being raised whether the environmental impacts caused by transportation and other factors justify the value of traditional paper presentations and social events. There is a pressing need to critically evaluate whether the ecological consequences of these events outweigh their perceived benefits. To that extent, we conducted a questionnaire survey among participants of the 45th International Conference on Software Engineering (ICSE) 2023 in Melbourne, Australia, seeking their feedback on the different conference sessions (e.g., workshops, keynotes, paper presentations, social events). In total, 161 participants filled out our survey. Overall, the conference was rated with 4.4 stars out of 5 stars. We do not see any significant differences among the different sessions, making it difficult to derive conclusions about their certain value and implications to sustainability. The relatively low response rate (11%) did not help in gaining better insights. Based on the participants registration data, we additionally estimated the carbon footprint emerged from air travel. The total carbon dioxide
The carbon footprint of academic conferences becomes a topic of increasing debate. It is important to consider whether the benefits derived from attending conferences in person outweigh the community's carbon footprint. Therefore, we need to evaluate the overall ecological consequences in relation to the perceived advantages. To that extent, we conducted a post-conference questionnaire survey among participants of the 44th International Conference on Software Engineering (ICSE) 2022 in Pittsburgh, USA, seeking their feedback about the conference and experience from a sustainability perspective. In total, 53 participants filled out our survey. Overall, 8 of 42 respondents felt that the community's carbon footprint was not offset by the benefits of in-person attendance.