arXiv:2607.12391v1 Announce Type: new
Abstract: We present a diffusion based model for asynchronous time series prediction, where the goal is to predict the next inter event time and event type. To address the inherent uncertainty of future events, we introduce ReDiTT, a retrieval augmented conditional diffusion transformer that operates in latent space. ReDiTT retrieves structurally similar latent sequences from a memory bank during both training and inference and incorporates them as reference conditions through cross attention. This retrieval based conditioning allows the model to attend to relevant temporal dynamics and provides global structural guidance for generation. As a result, ReDiTT stabilizes long horizon forecasting and improves sample diversity. Experiments on seven real world datasets demonstrate state of the art performance on next event prediction and long horizon forecasting. Our code is available at https://github.com/BorealisAI/ReDiTT.
Science Journals
arXiv:2607.12721v1 Announce Type: new
Abstract: In this paper, we propose an inversion method based on the firefly particle swarm optimization (FA-PSO) algorithm to estimate the permittivity, conductivity, and thickness of building materials using the free-space method. To improve convergence efficiency and robustness, an adaptive firefly algorithm (FA) is employed to systematically optimize the hyperparameters of the particle swarm optimization (PSO). By optimizing the parameters of the Gaussian distribution used for population initialization, the accuracy of parameter estimation is gradually improved. Furthermore, we derive the Cramer-Rao lower bound (CRLB) for the permittivity, conductivity, and thickness under a complex Gaussian noise model, which serves as a theoretical benchmark for evaluating the estimation accuracy of the FA-PSO algorithm. Numerical results indicate that for relatively thin materials, the estimation accuracy of the proposed method approaches this theoretical lower bound, confirming the effectiveness of the inversion framework. This study accurately extracts the electromagnetic properties of building materials, providing strong support for evaluating their wireless performance.
arXiv:2607.12803v1 Announce Type: new
Abstract: This paper presents a least squares weak Galerkin (LS-WG) finite element method for a class of second order elliptic equations of Fokker-Planck type. To address the numerical challenges arising from non-smooth diffusion tensors, the proposed method utilizes a least-squares formulation that yields a symmetric positive definite (SPD) discrete system. The numerical scheme is designed by employing locally constructed weak second order partial derivatives and the weak divergence commonly used within the weak Galerkin framework. A rigorous theoretical foundation is provided, establishing the uniqueness of the discrete solution and deriving optimal-order error estimates in a discrete energy norm. Finally, extensive numerical experiments are reported to validate the theoretical findings and demonstrate the robustness and performance of the numerical scheme.
arXiv:2607.12858v1 Announce Type: new
Abstract: Accurate open-world obstacle detection is critical for autonomous driving. Current anomaly segmentation methods suffer from a fundamental blind spot: they over-rely on texture novelty to identify out-of-distribution (OoD) objects while ignoring contextual spatial logic. Furthermore, mitigating the resulting false positives often requires cascading massive vision models, introducing unacceptable inference latency. To address these issues, we propose Layout-Aware Road Anomaly Detection (LARAD), shifting the paradigm from appearance matching to spatial-logic reasoning. First, we introduce the Spatial-Logic Violation Synthesis (SLVS) pipeline, which generates training samples that are texture-consistent yet spatially invalid, forcing the model to learn contextual violations. Second, we augment a standard closed-set segmentation network with a lightweight, OoD-guided attention branch. Extensive experiments demonstrate that LARAD significantly enhances robustness against logical anomalies and establishes a new state-of-the-art, all while retaining the high efficiency of a single-model architecture.
arXiv:2607.12866v1 Announce Type: new
Abstract: Reconstruction-based methods are a cornerstone of unsupervised image anomaly detection, but they remain vulnerable to \emph{outlier leakage}, where standard mean squared error (MSE) loss drives the model to faithfully reconstruct anomalous patterns. We propose a Non-linear Reconstruction Loss that applies a sigmoid-based squashing function to suppress high-magnitude features, preventing outliers from dominating optimization while preserving sensitivity to normal patterns. In addition, we introduce a statistical calibration scheme that selects the scaling factor $k$ from the confidence interval (CI) of the normal feature distribution, enabling data-driven control of the suppression strength. Our approach achieves competitive or superior anomaly detection performance compared to state-of-the-art methods, reaching 99.0\% Image-AUROC and 97.3\% Pixel-AUROC on MVTec-AD, and 95.3\% Image-AUROC and 99.0\% Pixel-AUROC on VisA. These results indicate that non-linear gradient suppression is an effective mechanism for mitigating outlier leakage and improving anomaly localization in unified industrial inspection settings. The implementation is available at https://github.com/mintii13/Statistical-Non-linear-Reconstruction-Loss.git.
arXiv:2607.12576v1 Announce Type: new
Abstract: Anomalous Sound Detection (ASD) aims to determine whether faults have occurred by monitoring sounds. Existing methods detect a limited range of anomalies, exhibit poor generalization, or train a separate model for each machine. Diffusion models possess strong generalization and can generate specific data with condition guidance. We propose a unified diffusion model only with a small module. The audio is first transformed into log-Mel spectrograms. The lightweight module embeds machine IDs into condition embeddings, guiding the model to reconstruct data for specific machines. Then diffusion model reconstructs data with condition, using Gaussian Mixture Models to fit the distributions of reconstruction errors. Our unified model could monitor multiple machine types and learn more fundamental feature spaces with cross-domain learning. Experiments on DCASE2022 Challenge Task 2 show that our model achieves 3.44% AUC and 2.52% pAUC improvements over baseline, validating its effectiveness.
arXiv:2607.12090v1 Announce Type: cross
Abstract: An independent set in a graph $G$ is a set of pairwise non-adjacent vertices. A tree decomposition of $G$ is a pair $(T, \chi)$ where $T$ is a tree and $\chi : V(T) \rightarrow 2^{V(G)}$ is a function satisfying two axioms: for every edge $uv \in E(G)$ there is an $x \in V(T)$ such that $\{u,v\} \subseteq \chi(x)$, and for every vertex $u \in V(G)$ the set $\{x \in V(T) | u \in \chi(x)\}$ induces a non-empty and connected subtree of $T$. The sets $\chi(x)$ for $x \in V(T)$ are called the bags of the tree decomposition. The tree-independence number of $G$ is the minimum taken over all tree decompositions of $G$ of the maximum size of an independent set of the graph induced by a bag of the decomposition. A graph $H$ is an induced minor of a graph $G$ if a graph isomorphic to $H$ can be obtained from $G$ by vertex deletions and edge contractions.
We prove that for every $t\in\mathbb{N}$ there exists an $\epsilon > 0$ such that every graph $G$ either contains the complete bipartite graph $K_{t,t}$ or the wall $W_{t\times t}$ as an induced minor, or has tree-independence at most $O(2^{O((\log n)^{1-\epsilon})})$. This leads to algorithms with running time $2^{n^{o(1)}}$, for a wide range of problems on $\{K_{t,t}, W_{t\times t}\}$-induced minor free graphs. Our result is a substantial generalization of existing bounds for the tree-independence and tree-width on various graph classes, and a partial resolution of the conjecture of Chudnovsky, E S, and Lokshtanov [Arxiv, 2025] that $\{K_{t,t}, W_{t\times t}\}$-induced minor free graphs have poly-logarithmic tree independence number. The generality comes at the cost of a sub-polynomial, rather than poly-logarithmic upper bound. Our result leads to a complete classification of induced-minor closed classes into ones that have sub-polynomial tree-independence, tree-independence equal to $\tilde{O}(\sqrt{n})$, and linear tree-independence.
arXiv:2607.12821v1 Announce Type: cross
Abstract: This study investigates the use of quantum annealing for particle track reconstruction in strip-type gaseous detectors. In such detectors, ghost hits and multiple hit combinations can turn pattern recognition into a combinatorial optimization problem. We formulate two reconstruction subproblems as quadratic unconstrained binary optimization problems. The first subproblem selects detector hits associated with a single photon track inside a localized candidate region. The second subproblem selects cluster triplets from different detector layers so that multiple track candidates can be handled within a single quantum processing unit(QPU) submission. The proposed formulations are tested using simulated DAMSA detector events. For the single track hit selection task, the QPU based reconstruction gives position and angular resolutions close to those obtained with a Kalman based reconstruction. In the simultaneous association task, valid cluster triplets are first extracted from the QPU samples and then connected using an association rule based on graph connectivity to construct track candidates. The DAMSA event topology studied here has low pileup and is dominated by the two photon signal from axion-like particle(ALP) decay. In this setting, the results show that the QUBO formulations can reproduce local reconstruction decisions. This provides a practical basis for further studies of reconstruction methods that combine quantum and classical computing in more complex tracking environments.
arXiv:2607.12983v1 Announce Type: cross
Abstract: Unlike in finite dimensions, quantum information in continuous-variable systems has the peculiar feature that without imposing physical constraints, the sample complexity of state tomography can be unbounded. Remarkably, this is even the case for state-of-the-art protocols for learning Gaussian states, which have finite-dimensional descriptions: the best known rates scale with $\log \log E$, where $E$ is the energy of the system. We prove this is not an artifact of existing analyses, but a fundamental limitation of the measurements used. We show: (1) Any protocol that uses Gaussian measurements, even entangled or adaptively chosen ones, must incur a $\log \log E$ dependence. This answers an open question posed by a number of previous works. (2) There is a smooth tradeoff between the number of rounds of adaptivity and the energy dependence, and we give a matching protocol achieving this interpolated rate. (3) With highly entangled, non-Gaussian measurements, one can learn $n$-mode pure Gaussian states with $O(n^2 / \epsilon^2)$ samples, independent of $E$. This answers an open question posed by Chen et al. (4) A simple protocol based on the single-copy canonical phase POVM of Holevo and Helstrom learns single-mode pure Gaussian states with $O(1/\epsilon^2)$ samples, again independent of $E$.
Our results clarify the role of energy in bosonic state tomography and shed new light on the intriguing interplay between adaptivity, entanglement, and magic in quantum learning.
arXiv:2607.12125v1 Announce Type: new
Abstract: Between 2023 and 2026, frontier AI systems crossed documented human expert baselines on a growing set of bounded, well-specified, evaluable cognitive tasks, including graduate-level science questions, competition mathematics, software-engineering benchmarks, and structured diagnostic reasoning, while the length of tasks such systems can complete at 50% reliability doubled roughly every seven months. These crossings are rapid and broad, but the frontier is jagged: humans retain decisive advantages in long-horizon reliability, genuinely novel problems, calibrated self-knowledge, sample-efficient learning, and embodied action, and benchmark results overstate deployed capability for reasons that are themselves now documented, namely contamination, construct validity, vendor self-evaluation, and the gap between 50% reliability and the reliability that economic work requires. Concurrently, humans increasingly use these systems as cognitive extensions. The offloading literature predicts costs to unaided skill, and early field evidence is consistent with such costs, though the largest meta-analytic evidence on prior technologies points the other way, and the question of whether generative AI differs is open. Finally, the experimental record on human-AI collaboration shows that naive combination often underperforms the stronger partner, implying that the human contribution must be repositioned toward specification, verification, and oversight, a shift visible in experiments but, so far, barely visible in field labor-market data. This paper states the resulting position, rapid crossings on a jagged frontier with a human role that must be redesigned rather than defended, and draws out its theoretical and practical implications.
arXiv:2607.12243v1 Announce Type: new
Abstract: Extended Dynamic Mode Decomposition (EDMD) approximates Koopman operators from data, but a single global operator is inefficient when different state-space regions exhibit distinct local dynamics. We introduce Cluster-Weighted EDMD (CW-EDMD), which jointly learns a soft phase-space partition and a per-cluster EDMD operator. Its Expectation-Maximization (EM) objective assigns each transition based on both geometric proximity and prediction residuals, so clusters specialize where local Koopman models are accurate rather than where the data are dense. On Lorenz, damped pendulum, and Duffing systems, across 36 configurations and 10 seeds, CW-EDMD improves matched-degree EDMD in one-step and 5s-rollout prediction. Across 288 paired comparisons, there are significant error reductions in 258 cases, increases in 4, and no differences in 26. Median one-step error reductions are 57x, 2.7x, and 12x on pendulum, Duffing, and Lorenz, respectively.
arXiv:2105.12663v5 Announce Type: replace
Abstract: The diversity of communication paths in a network, especially non-minimal paths, is a key enabler of performance at extreme scales. We present EvalNet, a toolchain for scalable generation and analysis of over 25 important network topologies, such as Slim Fly, PolarFly, and Orthogonal Fat Trees, with a strong focus on path diversity metrics. EvalNet provides an extensive and fine-grained analysis of shortest and non-shortest paths, including their multiplicities, lengths, and interference. It supports exact measurement and visualization of bandwidth and throughput between every router pair, enabling unprecedented insight into routing potential. EvalNet also includes detailed models for construction cost and power consumption, and interfaces seamlessly with established simulators, which we tune to support large-scale evaluations on low-cost hardware. Using EvalNet, we deliver the widest and most comprehensive path diversity study to date, demonstrating how path diversity underpins throughput and scalability, and facilitating progress towards new frontiers in extreme-scale network design.
arXiv:2607.12917v1 Announce Type: new
Abstract: This volume contains the ten peer-reviewed papers presented at HLPP 2026, the 19th International Symposium on High-Level Parallel Programming and Applications, held on 9-10 July 2026 at the Institut Henri Poincare in Paris, France. The symposium covers high-level approaches to parallel programming: programming models, languages, libraries, algorithmic skeletons, compilers, and runtime systems for multi-core, GPU, and distributed platforms. The 2026 edition extended this scope towards artificial intelligence, with new topics on the parallel programming and performance of AI systems and on AI-assisted generation of parallel code, and opened with a dedicated session on parallelism and AI. Papers were selected through a single-blind review process, with three Program Committee reviews per submission, and appear in the order of the symposium program. Edited by Chong Li, Corinne Ancourt, and Gaetan Hains.
arXiv:2406.08493v3 Announce Type: replace
Abstract: The concept of {\em countable sets} is attributed to Georg Cantor, who established the distinction between countable and uncountable sets in 1874. The concept of {\em computable sets} emerged in the 1930s through the foundational work on computing models by \Godel, Church, and Turing. In this paper, we investigate the connection between countability and computability. A {\em counting bijection} of a set $S$ is a bijection from the set of natural numbers to $S$. We say $S$ is {\em enumerable} if it is either finite or admitting a computable counting bijection. Our initial investigation shows that a set $S$ is enumerable if and only if it is computable. This equivalence offers new insights into set theory and computability theory. We further show that a set is countable if and only if it admits a {\em counting order}, which is a well order satisfying the {\em proximal} property. Based on this concept, we provide a procedure whose existence gives a necessary and sufficient condition for a set to be countable. This procedure is an algorithm if and only if the set is computable. A counting bijection $f$ is {\em increasing} if $f(x)>f(y)$ whenever $x>y$. We prove that an infinite set $S$ of natural numbers is definable in first-order arithmetic if and only if $S$ has an increasing counting bijection. This result has a significant implication: the standard proof that every set $S$ of natural numbers is countable is invalid. This is because the existing proof establishes that $S$ has an increasing counting bijection, which (by our result) would imply that $S$ is definable in first-order arithmetic. This leads to a contradiction with Tarski's undefinability theorem when $S$ is the set of \Godel\ numbers of the true arithmetic sentences.
arXiv:2407.17610v2 Announce Type: replace
Abstract: We propose In-like Pr10+ as a candidate for the development of a high-accuracy optical clock with high sensitivity to a time variation of the fine-structure constant, (\dot alpha}/alpha, as well as favorable experimental systematics. We calculate its low-lying energy levels by combining the configuration interaction and the coupled cluster method, achieving uncertainties as low as 0.1%, and improving previous work. We benchmark these results by comparing our calculations for the (5s^2 5p 2P_1/2) - (5s^2 5p 2P_3/2) transition in Pr10+ with a dedicated measurement and for Pr9+ with a recent experiment, respectively. In addition, we report calculated hyperfine-structure constants for the clock and logic states in Pr10+.
arXiv:2409.10094v3 Announce Type: replace
Abstract: Out-of-Distribution (OoD) detection aims to justify whether a given sample is from the training distribution of the classifier-under-protection, i.e., In-Distribution (InD), or from an unknown out distribution. Recent researches have leveraged Diffusion Models (DMs) for OoD detection due to their powerful distribution modeling capability. Given an input image, an InD-pretrained DM produces a corresponding InD-aligned counterpart, which serves as a generative reference for comparison. However, existing DM-based methods typically assess this underlying discrepancy through visual-level distances in the raw image space, which may be misaligned with the distributional discrepancy relevant to OoD detection. In this work, we investigate the fundamentals of discrepancy assessment in DM-based OoD detection, asking how the discrepancy between an input and its DM-generated counterpart should be formulated, and in which representation spaces and with which metrics it should be measured. To this end, we propose to assess the discrepancy in a classifier-relative manner by exploiting the representation spaces of the classifier-under-protection, whose training on InD data encodes rich task-relevant InD knowledge. In particular, we quantify two types of discrepancy: feature-level covariate discrepancy in deep feature representations and logit-level concept discrepancy in output logits, enabling effective differentiation between InD and OoD samples. Moreover, a subspace-based strategy is devised to refine representations of the DM generation to promote discrepancy assessment. Together, these designs form our novel detection framework, namely DDR. Extensive experiments on the challenging large-scale ImageNet-1K dataset demonstrate the superior detection performance of DDR over both DM-based and non-DM-based methods.
arXiv:2607.12545v1 Announce Type: new
Abstract: Adversarial robustness research has produced hundreds of defended models over the past decade, yet the literature almost universally reports robustness results in isolation: standard (clean) accuracy and adversarial accuracy of the robust model are shown, but the gap to the corresponding vanilla model is rarely quantified. We introduce VanillaBench, a systematic benchmark that makes this gap explicit. For every adversarially-trained model catalogued by RobustBench across four threat models, we compute the accuracy difference against multiple vanilla references from Papers with Code, computed over both all entries and no-extra-data entries, the best vanilla model as of the robust model's publication year, and an architecture-matched baseline. Across all 186 robust models, the mean delta clean relative to the best vanilla model ranges from -7.7 to -29.5 percentage points, and even the single most robust model per track still trails its temporal vanilla counterpart by 4.0-21.0 points. The architecture-matched comparison, which isolates the effect of adversarial training from architectural differences, reveals a mean gap of -3.5 to -17.5 points. Restricting this architecture-matched comparison to models whose vanilla accuracy is known for the exact same architecture, rather than approximated from a related one, narrows the gap to -4.0 to -14.0 points. These results demonstrate that the robustness-accuracy trade-off is substantially larger than what is typically conveyed by individual papers. This information is critical for practitioners and decision-makers. When deploying models in real-world settings, the accuracy cost of robustness directly affects business outcomes, yet current publications do not provide the vanilla baseline needed to assess it. We argue that future robustness evaluations should report vanilla-referenced accuracy gaps as a standard component.
arXiv:2607.12548v1 Announce Type: new
Abstract: We study the online preemptive matching problem, in which the edges of a graph arrive sequentially and the algorithm must maintain a matching by accepting or rejecting arriving edges and possibly discarding previously accepted ones. We prove a new upper bound of $0.5661$ on the competitive ratio achievable for the problem. This bound applies to arbitrary randomized algorithms, bipartite graphs and if we allow the algorithm to output a fractional solution.
Our result improves upon the strongest previously known upper bound of $2-\sqrt{2} \approx 0.585$, due to Huang et al. [SODA'19]. Previous hardness constructions relied on edge sequences described by vertex arrivals where each arriving vertex reveals its edges to yet unvaried vertices. Under such sequences, Huang et al. showed that there exists a non-preemptive online algorithm with competitive ratio $\sim0.567$ (or $2-\sqrt{2}$ for fractional solutions). Consequently, our hardness construction is the first result which shows hardness for instances where the optimal algorithm employs preemption.
arXiv:2603.14223v2 Announce Type: replace
Abstract: This work addresses an inverse reconstruction task for a time-fractional pseudo-parabolic model with a temporally varying coefficient. By imposing Dirichlet boundary conditions, we aim to recover the unknown initial state from observations collected at the final time.
From a theoretical perspective, we derive existence and uniqueness results by proving that, under suitable hypotheses, the problem admits a unique solution. Computationally, we introduce a finite-difference discretisation based on a time-stepping strategy and provide a detailed stability and convergence analysis. Leveraging the resulting forward solver, we then formulate an initial-data identification procedure using Tikhonov regularisation. The proposed approach is validated with numerical simulations, and its resilience is assessed via experiments that incorporate perturbations in the final-time measurements.
arXiv:2607.12569v1 Announce Type: new
Abstract: Fake speech detectors are increasingly challenged by the development of new and more accurate generative models. To cope with this problem, continual learning techniques are nowadays widely considered feasible strategies for updating models to new datasets, but they also lead to decreased performance on previously seen samples (catastrophic forgetting). In this work, we propose a forgetting-resilient solution based on the adoption of domain translators within a frozen detector, which remaps the new feature spaces into the original ones by means of a traceback translator network. Experimental results show that this strategy enables the achievement of high detection rates with respect to traditional retraining, while minimizing the computational effort and preserving the detection accuracy on previous data.
arXiv:2607.12826v1 Announce Type: new
Abstract: This paper addresses Carl Hempel's longstanding problem of statistical ambiguity in inductive-statistical inference, in which contradictory predictions are derived from statistical laws. To avoid such predictions, Carl Hempel proposed the Requirement of Maximal Specificity (RMS) for the statistical laws used in the inference. An analysis of the RMS refinements made by Wesley Salmon, Alberto Coffa, and James Fetzer led to the following definition of maximally specific statistical laws: "the lawlike premises of an adequate explanation must specify all and only those properties whose presence or absence made a difference to the occurrence of its explanandum-phenomenon." However, there was no proof of a solution to the statistical ambiguity problem based on this definition. We use Nancy Cartwright's definition of causes that raise probabilities across background contexts, and then introduce the concept of Causal Rules. Then we define a special semantic probabilistic inference procedure that incrementally refines these causal rules by incorporating all statistically relevant information. This procedure yields Maximally Specific Causal Relationships (MSCRs), for which we prove (Theorem 1) that predictions derived from them are consistent. This resolves the statistical ambiguity problem. The semantic probabilistic inference procedure provides a probabilistic causal learning system, which may be used in such new areas as Causal AI and Causal Machine Learning. They fundamentally explore causal inference as a tool for understanding cause-and-effect relationships within complex systems. Properties similar to RMS remain under discussion. Several notions related to RMS are considered: invariant feature learning, invariant causal prediction, and spurious association.
arXiv:2409.16808v3 Announce Type: replace
Abstract: Modern applications such as autonomous vehicles, intelligent surveillance, and smart city systems increasingly require object detection on resource-constrained edge devices. Yet, there is still limited understanding of how different object detection models behave across heterogeneous edge devices and under varying scene complexity. In this paper, we benchmark YOLOv8 (Nano, Small, Medium), EfficientDet Lite (Lite0, Lite1, Lite2), and SSD (SSD MobileNet V1, SSDLite MobileDet) on Raspberry Pi 3, 4, 5 with/without Coral TPU accelerators, Raspberry Pi 5 with AI HAT+, Jetson Nano, and Jetson Orin Nano. We evaluate energy consumption, inference time, and accuracy, and further examine how accuracy changes with the number of objects in the input image. The results reveal clear trade-offs among accuracy, latency, and energy efficiency across model-device combinations. SSD MobileNet V1 achieves the lowest latency and energy consumption but the lowest accuracy, whereas YOLOv8 Medium achieves the highest accuracy at higher computational cost. TPU-based Raspberry Pi devices improve the efficiency of SSD and EfficientDet Lite while reducing YOLOv8 accuracy. Orin Nano offers the most favorable overall balance across most model families. The object-count-based analysis further shows that models achieve more similar accuracy on simpler images, while the accuracy gap widens as scene complexity increases.
arXiv:2411.02989v2 Announce Type: replace
Abstract: How diverse are the outputs of large language models when diversity is desired? We examine the diversity of responses of several language models to questions with multiple possible answers, comparing them with human responses. Our findings suggest that models' responses are highly concentrated, reflecting narrow, mainstream outputs, in comparison to humans, whose responses exhibit a much longer-tail. We examine three simple and practical ways to increase output diversity: 1) increasing generation randomness via temperature sampling; 2) prompting models to answer from diverse perspectives using a single prompt; 3) aggregating outputs from several models. We find that these interventions, especially when combined, can substantially increase output diversity, although single-model outputs generally remain less diverse than the human baseline. We discuss potential implications of these findings for future work in AI policy and governance that wishes to preserve cultural diversity, an essential building block of a democratic social fabric.
arXiv:2607.12113v1 Announce Type: new
Abstract: One year ago, the AISLE roadmap argued that autonomous laboratories operated as isolated islands and proposed a grassroots network organized around five critical dimensions. The field has since moved faster than anticipated. Multi-agent systems have produced experimentally validated hypotheses, self-driving laboratories have grown more interoperable and orchestrated, reasoning-trained and domain foundation models have raised the capability ceiling, and the Genesis Mission has placed autonomous experimentation at the center of U.S. federal science strategy, with industry emerging as a primary actor. Progress has met a sobering counter-current, including a corrected flagship discovery result, benchmarks showing that agents which rival experts on closed-ended questions still complete only a fraction of open-ended research, and fabricated citations surfacing at leading venues. We read this as the defining tension of the field. Producing a candidate discovery is no longer the hard part, but verifying it is, and this asymmetry now limits autonomous science more than raw model capability. We update the roadmap around seven dimensions, revisiting the original five and elevating two former cross-cutting concerns, trust, verification, and reproducibility, and safety, security, and governance, to first-class status. We assess the original milestones (M1 through M14) as achieved, partially achieved, reframed, or open, add four new milestones (M15 through M18), and scope the path forward to a two-year horizon. The first year concentrates on interfaces, protocol adoption, and the scaffolding of verification, and the second targets federation, zero-trust coordination, and governance. Throughout, we position the grassroots network as the interoperability fabric that lets national programs, international initiatives, and commercial platforms connect rather than re-silo.
arXiv:2411.06608v3 Announce Type: replace
Abstract: We introduce MolMiner, a fragment-based, geometry-aware, and order-agnostic autoregressive model for molecular design. MolMiner supports high-dimensional conditional control over twelve physicochemical and structural properties from partial specifications, constructs molecules via symmetry-aware fragment attachments, and conditions each generation step on force-field-relaxed three-dimensional geometry of the partial structure. Conditional control emerges without auxiliary property losses. On targeted property windows, conditioning lifts hit rates by up to 5.25x over unconditional generation and 3.5x over the training distribution itself -- overriding the model's intrinsic biases -- at the cost of a small reduction in unconditional distributional fidelity. MolMiner unifies dynamic geometry, symmetry handling, order-agnostic generation, and scalable multi-property conditioning within a single framework.