Peer Revenue Now accepting positions

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Thursday 8 October

Machine Learning

Prospective Prediction of OOD Degradation from Source-Side Training Dynamics

Sasha Monin

We study whether persistent out-of-distribution (OOD) degradation can be predicted before it is directly observed using only source-side training dynamics. In a controlled shortcut-learning setting, a simple logistic regression predictor develops a clear prospective signal, while training time alone does not. Temporal summaries of the source-side quantities are substantially more informative than their current values. When transferred without additional training from a CNN to an MLP, confidence and entropy dynamics retain substantial predictive information. These results provide a proof of principle that source-side training dynamics can contain an early warning signal for future OOD failure.
Read the full abstract

Statistics Applications

Mixup Barcodes for Topology-Aware Financial Decision Making

Buddha Nath Sharma, Joe Opitz, Adam Moser et al. (+3 more)

We present a topology-aware system based on mixup barcodes and persistent homology for financial decision making. The suggested approach uses topological summaries obtained from Takens delay embeddings of a univariate price series to quantify structural changes between reference and current market regimes. The 1-Wasserstein distance between persistence diagrams, a mixup barcode disruption index, and persistence entropy divergence are combined to provide a novel stress score. A Golden Cross trading method is dynamically modulated by this score, which transforms a binary buy/sell signal into a continuous position-sizing process. At embedding settings chosen by maximizing in-sample Sharpe over a 72-point grid, in-sample assessment on the S&P 500 and Bitcoin yields Sharpe ratios of 1.116 and 1.238, respectively, with much lower maximum drawdown compared with both Buy-and-Hold and the standard Golden Cross baseline. Our findings imply that topological summaries of market geometry include useful information that goes beyond traditional trend markers.
Read the full abstract

All authors Buddha Nath Sharma, Joe Opitz, Adam Moser, Sayam Palrecha, Sushovan Majhi, Atish Mitra

Numerical Analysis

A neural characteristic mapping method: Lagrangian PINNs based on flow maps for transport-dominated problems

Martin Campos Pinto, Victor Fournet, Emmanuel Franck et al. (+2 more)

We propose a novel numerical method for transport-dominated problems using Physics-Informed Neural Networks (PINNs). In such problems, the solution can develop large gradients and fine structures over time. This is particularly challenging for classical PINNs, which represent the solution itself. In our approach, rather than the solution, we choose to approximate the flow map of the equation. The solution is then recovered by pulling back the exact initial condition, so that its fine structures are produced by the map rather than represented by the network. Our method is constructed by approximating the flow map of the equation with multiple PINNs, one per time subinterval, and composing them using the semigroup property of the flow, so that each network only has to represent a map close to the identity. To highlight the advantages of our method compared to classical PINNs or classical numerical schemes, we present numerical results on the linear advection equation, the incompressible Euler equation in vorticity formulation, the Vlasov-Poisson equation, and finally a drift-kinetic equation.
Read the full abstract

All authors Martin Campos Pinto, Victor Fournet, Emmanuel Franck, Philipp Krah, Victor Michel-Dansac

Astrophysics of Galaxies

Inflow-driven galaxy evolution - II: A hierarchy in the scaling relations of star-forming galaxies

Kai Wang, Carlton Baugh, Sownak Bose et al. (+7 more)

The star-forming main sequence (SFMS; $M_\star$–sSFR), the gas-to-stellar mass relation ($M_\star$–$M_{\rm gas}/M_\star$), and the mass–metallicity relation (MZR; $M_\star$–$Z_{\rm gas}$) are three core scaling relations for star-forming galaxies, and together they encode the information about the baryon cycle that operates during galaxy evolution. We investigate how these relations emerge by modelling the gas flows that regulate the baryon cycle, using both an ideal and a cosmological gas flow model, and identify a hierarchy among the three relations, with one further parameter entering the scatter budget at each step. In our model, the SFMS and its scatter are set primarily by the formation time of the host halo, with a weak dependence on the star formation efficiency and the mass-loading factor. The gas-to-stellar mass relation follows from the ratio of the sSFR to the star formation efficiency, and inherits the scatter of both. The MZR in turn follows from the gas-to-stellar mass relation together with the mass-loading factor, their relative contributions depending on the galaxy evolutionary state between the inflow-driven regime and equilibrium. This hierarchy naturally introduces further correlations among the three relations, such as the fundamental plane of star formation ($M_\star$–$M_{\rm gas}$–SFR), the fundamental metallicity relation ($M_\star$–SFR–$Z_{\rm gas}$) and its gaseous counterpart ($M_\star$–$M_{\rm gas}$–$Z_{\rm gas}$), and the correlations of each relation with galaxy size, which traces the star formation efficiency through the star formation law. Together, the three core scaling relations and the higher-dimensional correlations among them follow from a single framework, providing a foundation for linking the observables of star-forming galaxies to the underlying baryon cycle.
Read the full abstract

All authors Kai Wang, Carlton Baugh, Sownak Bose, Shaun Cole, Carlos S. Frenk, Cedric Lacey, Julio F. Navarro, Kyle A. Oman, Amélie Saintonge, Tom Theuns

Artificial Intelligence

HRIL: Learning Multimodal Synergy via Higher-Order Tensor Modeling

Qun Dai, Liangjian Wen, Jiang Duan et al. (+7 more)

Self-supervised multimodal representation learning has achieved remarkable success across diverse domains, yet capturing synergistic information remains challenging due to the complexity of cross-modal interactions. Unlike the shared information across individual modalities, synergy arises when task-relevant signals emerge only from the joint configuration of multiple modalities and cannot be recovered from any modality in isolation. This work focuses on how to preserve the information capacity for such synergistic signals in multimodal representations. The key observation is that synergistic information is reflected in higher-order statistical dependence among modalities, which provides a principled target for explicitly modeling joint interactions. Motivated by this insight, we propose Higher-order Representation and Information Learning (HRIL), which constructs an empirical cross-moment tensor over modality embeddings to represent multi-way interactions. HRIL employs Tucker decomposition to obtain a core tensor, complemented by a synergy-aware regularizer that prevents energy concentration and preserves higher-order coupling capacity for synergistic information capture. Experiments on the controlled synergy task and real-world benchmarks demonstrate consistent improvements over existing multimodal contrastive methods, with notable gains on tasks dominated by synergistic interactions. Code is released at https://github.com/brightest66/HRIL.
Read the full abstract

All authors Qun Dai, Liangjian Wen, Jiang Duan, Yong Dai, Dongkai Wang, Maolin Wang, Mingjie Wang, Jianzhuang Liu, He Yan, Zhao Kang

Computational Geometry

Tiling 3D by Translates of a Single Polycube is Undecidable

Erik D. Demaine and Stefan Langerman

We prove co-RE-completeness, and thus undecidability, of the following problem: given a single (connected) polycube, decide whether it tiles 3D Euclidean space by translations. We reduce from Wang tiling using the decorated two-prime Sudoku construction of Greenfeld and Tao and a cyclic encoding adapted from OpenAI's 3D aperiodic tile, and apply a reduction of Kim to make the prototile connected (via faces). Dimension three is optimal: translational monotiling is known to be decidable in $\mathbb{Z}^2$ and for a single (possibly disconnected) polyomino in $\mathbb{R}^2$.
Read the full abstract

Artificial Intelligence

GeoReform: Reflective Formalization Evolution for Multimodal Geometry Problem Solving

Jialu Wang, Ruichen Zhang, Xiaoou Liu et al. (+2 more)

Multimodal large language models (MLLMs) often struggle to identify and use geometric relations in diagrams. Recent methods address this challenge by converting geometric entities, relations, and constraints into explicit textual representations for the model to reason over. However, effective formalization is highly non-trivial: on Geometry3K, structure injection fixes 28 errors but introduces 13 new ones among 200 examples. Redundant relations can distract the model, while ambiguous references to diagram elements can lead it to apply constraints incorrectly. This suggests that the key challenge is not merely extracting more geometric facts, but organizing them into representations that support downstream reasoning. To fully exploit the power of formalization, we further propose GeoReform, a reflective formalization evolution framework that treats formalization as an optimizable policy rather than a fixed parser output. GeoReform executes the full reasoning pipeline, collects failed rollouts, diagnoses defects in the current representation, and mutates the policy to better select, ground, group, and present geometric entities, relations, constraints, and targets. On Geometry3K, GeoReform improves Qwen3VL-2B accuracy from 42.0% to 56.0%. Extensive experiments and analyses across geometry reasoning benchmarks demonstrate that effective formalization is crucial for improving multimodal geometry reasoning.
Read the full abstract

All authors Jialu Wang, Ruichen Zhang, Xiaoou Liu, Hua Wei, Tianlong Chen

Machine Learning

Long Text to Predictive Features: LLM-Guided Blockwise Feature Engineering via Executable Program Search

Ziming Dai, Dabiao Ma, Ziheng Guo et al. (+2 more)

Industrial risk-control systems typically rely on structured-data models for efficient prediction, yet substantial valuable information remains embedded in unstructured long text. Extracting this information through manual feature engineering is labor-intensive, while requiring a large language model (LLM) to process every real-time input may not meet practical deployment requirements. To address this challenge, we propose LLM-BlockFE, an LLM-guided offline feature construction framework that converts long text into executable feature programs, thereby avoiding LLM calls during online inference. LLM-BlockFE constructs feature programs by incrementally appending immutable code blocks and evaluates candidate features using a downstream model. To address the tendency of conventional greedy search to become trapped in suboptimal solutions, our method introduces a block-level rollback mechanism based on depth-calibrated credit allocation and advances multiple independent search trajectories in an interleaved manner, reducing redundant exploration by sharing fixed descriptions of each trajectory's exploration direction. After the search, the resulting programs are frozen and deployed to extract structured features for downstream prediction models. Across two public and two private datasets, LLM-BlockFE achieves absolute AUC improvements of 0.0069 to 0.0358 over the strongest baseline on each dataset in the full-dataset comparison. Post-launch monitoring across five deployed financial risk-control applications shows absolute KS improvements of 0.02 to 1.56 percentage points over the existing manually designed strategy.
Read the full abstract

All authors Ziming Dai, Dabiao Ma, Ziheng Guo, Jack Dong, Zimu Zhou

General Relativity and Quantum Cosmology

A Pairwise Surrogate for Gravitational-Wave Spectra from Highly Relativistic Vacuum Bubble Collisions

Malte Buschmann and Toby Opferkuch

Gravitational waves from vacuum first-order phase transitions probe the nonlinear dynamics of relativistic bubble collisions. Predicting their spectrum requires resolving Lorentz-contracted scalar bubble walls throughout a volume containing many bubbles, making the large-boost regime prohibitively expensive for direct lattice simulations. We develop a simulation-based pairwise surrogate that separates the microscopic collision dynamics from the geometry of the multi-bubble transition. This utilizes symmetry-reduced $(1+1)$-dimensional scalar-field simulations to build a library of two-bubble gravitational-wave spectra, including the nonlinear post-collision evolution. These spectra are combined with geometric weights that account for screening by other bubbles. We test the two-bubble spectra and the reconstructed spectra of three- and many-bubble systems against full $(3+1)$-dimensional lattice simulations. Not only can the surrogate be orders of magnitude cheaper computationally, but it also provides access to highly relativistic collisions beyond the practical reach of direct many-bubble simulations. This provides a route to studying how the spectral amplitude and shape depend on the collision boost and scalar potential.
Read the full abstract

Computer Vision

GenIA: Generative Reconstruction with Test-Time Input Alignment

Stefano Esposito, Naama Pearl, Polina Karpikova et al. (+5 more)

Reconstructing complete 3D object assets from monocular or sparse multi-view observations remains challenging. Generative 3D foundation models can complete object geometry beyond the observed views, but their predictions may not faithfully reproduce the observed geometry, appearance, or pose. We introduce GenIA, a framework for test-time input-aligned generation that grounds SAM3D's generative prior in geometric and photometric observations without retraining the foundation model. We improve object pose by deriving translation and scale from geometry while retaining the learned rotation prior, and align appearance through visibility-biased attention, cross-observation fusion, and differentiable rendering guidance during denoising. An optional post-denoising refinement further adapts the appearance latent, lightweight decoder adapters, and object placement to the observations. Our framework also supports externally supplied geometry; when given temporal shapes of dynamic objects, it recovers a shared, input-aligned canonical appearance and stable world-space placement. Across synthetic and real benchmarks, GenIA improves pose prediction and object reconstruction from monocular, multi-view, and dynamic inputs, outperforming recent optimization-based, per-frame image-to-3D, and video-to-4D methods. Our project page is available at https://facebookresearch.github.io/GenIA.
Read the full abstract

All authors Stefano Esposito, Naama Pearl, Polina Karpikova, Samuel Rota Bulò, Lorenzo Porzi, Peter Kontschieder, Andreas Geiger, Jonathon Luiten

Analysis of PDEs

Endpoint Compactness and Finite-Energy Weak Solutions for the 2D Isothermal Compressible Navier-Stokes

Cheng Yu

We prove global existence of finite-energy renormalized weak solutions to the 2D isothermal compressible Navier–Stokes equations in bounded $C^2$ domains with no-slip boundary conditions. The initial data are arbitrary in the natural energy class, allowing vacuum and unbounded density. This resolves the isothermal endpoint $\gamma=1$ of the classical 2D finite-energy renormalized existence theory for power-law pressures. We develop a new endpoint compactness argument based on a truncation-independent integrable bound for centered density oscillations. This argument yields strong $L^1$ compactness of the approximating densities at the natural energy level.
Read the full abstract

Robotics

ARC: A Reasoning Recipe for Robot Foundation Models

Gokul Puthumanaillam, Tao Sun, Elie Aljalbout et al. (+5 more)

The prevailing approach to improving robot foundation models (RFMs) relies on larger models, more robot demonstrations, and costly training at scale. We show that there exists an effective and efficient complementary approach: the right reasoning recipe can substantially improve the zero-shot task performance of existing state-of-the-art RFMs. We refer to this recipe as ARC. It consists of three key ingredients: a reasoning trace, a scalable automatic labeling pipeline, and a strategy for adapting pretrained RFMs to use these traces for control. First, we find that effective reasoning traces should be grounded in the robot's next action and explain its causal structure: why the action is appropriate and what effect it should produce. Second, we show that these traces can be generated automatically from existing demonstrations, enabling us to construct ARC-Trace-DROID from DROID without collecting new robot data. Third, we show how state-of-the-art VLAs such as $\pi_{0.5}$ and WAMs such as Cosmos3-Nano-Policy can learn to use these traces for control, with fine-tuning and inference tailored to each model's architecture and capabilities. Using ARC, we obtain gains in zero-shot RFM performance that, to our knowledge, are unprecedented without additional robot demonstrations or foundation-scale training. The adapted models establish a new state of the art on RoboLab-120 and MolmoSpaces, with gains of up to 50 percentage points on RoboLab-Reasoning-50. On real robots, ARC improves $\pi_{0.5}$'s task success by 82.2 percentage points. Project website: https://arc-robot-reasoning.github.io/
Read the full abstract

All authors Gokul Puthumanaillam, Tao Sun, Elie Aljalbout, Moritz Reuss, Zhaoshuo Li, Fabio Ramos, Ankit Goyal, Jenai Xuning Yang

High Energy Physics (Theory)

Platonic self-dual solitons in a (1+1)-dimensional $\mathbb{CP}^1$ model

L. A. Ferreira and G. Marcelino

We construct a family of self-dual $(1+1)$-dimensional $\mathbb{CP}^1$ sigma models whose potentials $V$ are determined from pre-potentials $U$ by Platonic geometry. Using the round metric on $\mathbb{S}^2\simeq\mathbb{CP}^1$ and the generalized self-duality formalism, we build symmetry-invariant pre-potentials $U$ from the unit vertices of each Platonic solid intersecting the unit circumsphere. The regular-lattice sites associated with vertex, face-center, and edge-midpoint directions are critical points of $U$ and degenerate vacua of the scalar potential $V$. Dual solids share the same regular lattice but generate different pre-potentials, charges, potentials, and metric-gradient flows. Numerical integration reveals distinct self-dual trajectories and energy-density profiles, including multi-peak structures for paths passing near saddle points, while solutions with the same endpoint difference $\Delta U$ have the same total energy. The positive maxima of $V$ are also organized into symmetry orbits. This construction provides a geometric prescription for controlling the vacuum structure and self-dual sectors of $\mathbb{CP}^1$ models with a potential.
Read the full abstract

Algebraic Geometry

Local Constancy of the Betti Automorphic Category over the Moduli of Curves

Joakim Færgeman and Marius Kjærsgaard

Nadler and Yun conjectured that the automorphic side of the Betti geometric Langlands correspondence forms a local system of categories over the moduli of curves. We prove this local constancy conjecture using purely microlocal methods, independent of the recently established Betti geometric Langlands conjecture. In particular, it follows that the Betti automorphic category depends only on the genus of the curve, verifying a conjecture of Ben-Zvi and Nadler.
Read the full abstract

High Energy Physics (Theory)

Differential $T_2$-Duality and Spans of Principal 3-Bundles with Connections

Gianni Gagliardo, Christian Saemann and Roberto Tellez-Dominguez

We present a generalization of T-duality, called $T_2$-duality, that captures certain aspects of U-duality in M-theory. In particular, it features backgrounds that contain a 2-gerbe over a principal torus bundle, structures familiar from eleven-dimensional supergravity backgrounds. In special cases, it also incorporates S-duality in a precise sense. Our main result is a theorem characterizing $T_2$-dual pairs through spans of categorified principal bundles with connections, giving explicit rules for constructing a $T_2$-dual for a given background. A number of explicit examples illustrate how the various characteristic classes are mapped between the left and right backgrounds. For two-dimensional torus fibers, we also relate our construction to the combination of dimensional reduction and T-duality that links eleven-dimensional supergravity to type IIB supergravity. Finally, we discuss how $T_2$-duality arises from a Buscher-like span of 3d sigma models.
Read the full abstract

Computer Vision

WorldAlign: Decoupled 4D Reward for World-Consistent Video Generation

Jing He, Kaixin Ding, Xingye Tian et al. (+5 more)

Faithful visual world simulation requires generated videos to maintain 4D world consistency, encompassing both static and dynamic consistency. Static consistency requires coherent 3D structure in static environments across viewpoints, while dynamic consistency requires plausible subject motion and consistent appearance over time. Geometry-aware post-training offers a promising way to improve world consistency. However, existing methods often rely on a static-scene assumption. Even those that accommodate dynamic scenes struggle to provide reliable static-consistency feedback, while dynamic consistency is often overlooked or inadequately assessed. To address these limitations, we introduce WorldAlign, a decoupled 4D reward framework that semantically separates static regions and dynamic subjects and provides feedback by aligning each with a world prior suited to its assumptions. For static regions, WorldAlign aligns static geometry with a geometric world prior through semantically guided masked reprojection, enabling more reliable static-consistency evaluation; an auxiliary camera-motion reward discourages nearly static solutions. For dynamic subjects, WorldAlign uses a strong vision-language model (VLM) as a dynamic world prior and constructs a VLM-as-a-judge reward based on sample-specific checklists that assess dynamicity, physical plausibility, shape, and texture consistency. This decoupled design enables more effective online post-training without requiring human preference annotations. Across two pretrained image-to-video generators, Wan2.1 and Wan2.2, WorldAlign jointly improves static and dynamic consistency over existing methods without suppressing overall or subject motion. These results support decoupled world-prior alignment for more faithful visual world simulation. Project page: https://worldalign.github.io/.
Read the full abstract

All authors Jing He, Kaixin Ding, Xingye Tian, Guibao Shen, Wenhang Ge, Xin Tao, Pengfei Wan, Ying-Cong Chen

Statistical Methodology

Data generation for the logistic regression model under controlled conditions

Kabiru Abubakari and Silvia Liverani

Generating data with desired properties for a logistic regression model under controlled conditions is challenging. We propose an integrated iterative bisection algorithm that determines values of the intercept and regression coefficients capable of generating data with a specified prevalence and predictive accuracy. The performance of the proposed method is demonstrated through simulation studies. We further show that the choice of parameter values, or, in a Bayesian setting, the choice of priors for these parameters, induces a particular level of predictive accuracy. This relationship should therefore be taken into account when generating data for model evaluation under controlled conditions.
Read the full abstract

Signal Processing

Interference-Aware Downlink Power Control for High-Altitude Platform Stations with QoS and Max-Min Fairness

Rajan Shrestha and Hayder Al-Hraishawi

High-altitude platform stations (HAPS) are a promising component of non-terrestrial networks (NTNs) for wide-area connectivity. However, serving multiple users over large user footprints requires efficient power-control mechanisms that account for the unique propagation characteristics and strict power constraints of HAPS. This paper develops a tractable analytical and optimization framework for downlink power allocation in HAPS systems under statistical channel state information, 3GPP NTN path loss, and Rician fading. A closed-form expression for the signal-to-interference-plus-noise ratio (SINR) with mean-channel maximal-ratio transmission (MRT) precoding is derived, capturing geometry-dependent inter-user interference coupling. Based on this, two power-control schemes are proposed: quality-of-service (QoS)-constrained power minimization and max-min SINR fairness. The QoS solution is obtained in closed form or through a low-complexity fixed-point iteration, while the max-min solution is obtained through bisection. Numerical results demonstrate up to 6 dB reduction in required transmit power at moderate-to-high target rates and improved worst-user SINR compared to the uniform power, inverse path-loss, and geometry-aware heuristic allocations. The proposed QoS-optimal scheme also achieves a QoS feasibility probability of approximately 0.95 at R_req = 2.5 bits/s/Hz and P_max = 35 dBm, while the uniform power, inverse path-loss, and geometry-aware heuristic allocations achieve less than 0.4. Moreover, the proposed schemes achieve performance comparable to the linear programming (LP)-based benchmarks at a fraction of the computational cost, reducing the average computation time by approximately 99.97% when K = 20.
Read the full abstract

Machine Learning

Marformer: A Transformer for Predicting Missing Data Distributions

Prabhav Singh, Xiheng Tom Wang, Haojun Shi et al. (+1 more)

Real decisions are made under incomplete information. If we observe only some of the random variables we need, we can predict the others. The conditional marginals over the missing variables are the key ingredient for computing Bayes risk and Value of Information (VOI), the expected gain from acquiring one more observation before deciding. We present the Marformer, a Transformer trained to directly predict conditional marginals given any set of observed values. Like BERT, which is trained to predict missing words from context, the Marformer constructs a hidden-vector representation for each distribution $p(X_i)$ and iteratively refines it through attention to other distributions $p(X_j)$. Unlike generative approaches, the Marformer does not model the full joint distribution, requires no domain knowledge of the data-generating process, and makes all predictions in a single forward pass. We evaluate across three synthetic domains with missing data—Bayesian networks, discretized multivariate Gaussians, and structured annotation data. The Marformer can match or outperform classical missing-data methods, even when those methods are given the true model family and prior that generated the synthetic data. We also evaluate on a real annotation dataset, where the Marformer outperforms the evaluated baselines at the largest training size. In both cases, the Marformer is substantially faster than the evaluated generative baselines.
Read the full abstract

All authors Prabhav Singh, Xiheng Tom Wang, Haojun Shi, Jason Eisner

Computational Complexity

On the Hardness of $4$-to-$1$ Games with Perfect Completeness

Yumou Fei, Dor Minzer and Shuo Wang

We prove that for all $\varepsilon>0$, there exists a positive integer $k$ such that given a $4$-to-$1$ game $\Psi$ with alphabet size at most $k$, it is $\mathbf{NP}$-hard to distinguish between the case that $\mathrm{val}(\Psi)=1$ and the case that $\mathrm{val}(\Psi)\leq \varepsilon$. This confirms the $4$-to-$1$ Games Conjecture from [Khot, CCC 2002]. Previously, the best known result, due to [Dinur, Khot, Kindler, Minzer, Safra], established the almost-perfect completeness version (but applied to the stricter problem of $2$-to-$1$ games). Using results from the literature, we get the following implications: (1) for all $k\in \mathbb{N}$, given a $3$-colorable graph $G$, it is $\mathbf{NP}$-hard to find a proper $k$-coloring; (2) for all $\delta>0$, given a $2$-colorable $3$-uniform hypergraph $G$, it is $\mathbf{NP}$-hard to find in it an independent set containing at least $\delta$ fraction of the vertices. Our proof is a three-step construction that builds on the two-step framework of [Dinur, Khot, Kindler, Minzer, Safra]. In the outer-PCP step, we use quadratic equations to gain perfect completeness. We then construct a new middle PCP that performs low-rank tests while preserving a key covering property. Finally, we construct a new inner PCP based on a tensor of the standard Grassmann encoding with its low-rank variant due to [Golowich, FOCS 2023].
Read the full abstract

Statistical Methodology

Formalization and consistency of clone-censor-weight

Iqraa Meah, Charlotte Voinot, Stefan Michiels et al. (+5 more)

Target trial emulation has become a standard framework for causal inference from observational data. Within this paradigm, the clone-censor-weight (CCW) methodology provides a practical way to deal with complex treatment and adherence definitions when treatment regimes are not distinguishable at baseline. Despite its increasing use, the statistical interpretation of CCW remains limited. In this work, we formalize the CCW methodology from a statistical viewpoint based on admissible sets and stochastic interventions. We characterize the causal estimands targeted by CCW, establish consistency of CCW estimators under standard identification assumptions, and provide theoretical guarantees for bootstrap-based inference methods.
Read the full abstract

All authors Iqraa Meah, Charlotte Voinot, Stefan Michiels, Noémie Simon-Tillaux, Claire Jouffroy-Zeller, Inès Vaz-Luis, Clément Berenfeld, François Petit

Computation and Language

Latent Core Tokenizer: Compress, but Meaningfully

Felermino D. M. A. Ali, Millicent Ochieng, Ogbemi Ekwejunor-Etchie et al. (+3 more)

Tokenizers are commonly optimized for compression, but a compact vocabulary does not necessarily distribute its capacity evenly across languages. We introduce the Latent Core Tokenizer (LCT), a language-agnostic approach that separates structural discovery from vocabulary construction. LCT uses Minimum Description Length, entropy-based boundary signals, and morphotactic constraints to identify reusable linguistic units before constructing a shared vocabulary. Across 104 languages with a 200K-token vocabulary, LCT achieves lower fertility and higher MorphScore than BPE, Unigram, and parity-aware BPE, while maintaining comparable cross-lingual disparity in tokenization cost. Across four multilingual downstream benchmarks, LCT improves aggregate score by 1.48, 1.83, and 2.00 points over BPE, Unigram, and parity-aware BPE, respectively. Our findings show that compression alone does not predict representation quality and highlight the importance of morphology-driven structural discovery and how frequency is used to allocate the final vocabulary across languages.
Read the full abstract

All authors Felermino D. M. A. Ali, Millicent Ochieng, Ogbemi Ekwejunor-Etchie, Ade Famoti, Jacki O'Neill, Debjit Paul

Artificial Intelligence

OnTrack: Real-Time Monitoring and Intervention in LLM Agent Trajectories via Streaming Structure-Aware Optimal Transport

Babak Barazandeh, Connor Swanson, Chinmay Kulkarni et al. (+1 more)

Agents are deployed in applications from trip planners and stock trading to IT incident triage. In most cases, LLM agents work autonomously with minimal rule-based safeguarding, leading to cost and safety issues from irreversible actions. Recent works resolve this either by using a safeguard agent to monitor behavior or evaluating logs post-hoc. The first adds cost and latency to every step; the second delivers its verdict after the run, when tokens are burned and damage is done. To overcome this, we propose OnTrack, a streaming monitoring mechanism that compares an agent's steps and dependencies against recorded successful runs to alert users or block the agent in about a millisecond per step. We study this problem in three regimes of decreasing access: full reference access (historical runs and tool schemas), intermediate access (only tool schemas), and no prior knowledge (only step logs as generated). Expectation of OnTrack's monitoring capabilities reduces as data access drops, ranging from plan violation detection to identifying loops, stalls, and repeated tool calls. Finally, we evaluate OnTrack using SWE-bench trajectories. Based on the first 8 steps, our method ranks failing trajectories below succeeding ones better than content similarity approaches (+0.057 AUROC). With an abort policy, we save about 18% of compute that would be burned on failing runs, where 83% of interrupted runs were actually heading to failure (5 out of 6 aborts were correct).
Read the full abstract

All authors Babak Barazandeh, Connor Swanson, Chinmay Kulkarni, Nikhil Mungel

Computer Vision

AgentGarten: Code Worlds for Evolving Agents

Jiawei Chi, Shangchen Miao, Zhiyuan Shi et al. (+12 more)

Interactive virtual worlds allow agents to learn through exploration and interaction. What agents can learn is bounded by the environments they practice in, which must be faithful, with consistent state, rules, and dynamics, and realistic, with observations that follow the real-world visual distributions. Achieving both across diverse worlds remains a bottleneck. We introduce AgentGarten, a framework that couples simulators and game engines with a shared neural renderer to build real-time interactive environments. Its simulation backends maintain persistent world state and execute program-defined interaction rules, while the renderer generates visual observations from structured conditions exported through a common interface. To build the neural renderer, we adapt a pretrained video model to geometry conditions, distill it with our proposed Adversarial Forcing, and optimize inference for real-time interaction. Adversarial Forcing makes history prefilling differentiable through exact replay, so that losses on later predictions update how the renderer encodes prior observations, and adds real-data adversarial supervision to improve its visual quality. In AgentGarten, agents perceive the world through visual observations, interact with it in real time, and improve by distilling each round of experience into playbooks that subsequent agents inherit and refine. Our empirical study demonstrates a substantial gain in learning efficiency, with agents learning from just 4 rounds compared with millions for a conventional reinforcement learning counterpart. As new worlds can be written as code and rendered through the same interface, environments can scale in both number and difficulty alongside their agents, a step toward agents that keep evolving through interactive experience.
Read the full abstract

All authors Jiawei Chi, Shangchen Miao, Zhiyuan Shi, Kailu Wu, Hanyang Wang, Weiliang Chen, Qiyu Dai, Chengyang Li, Jinshan Ren, Jun Gao, Mingsheng Long, Yueqi Duan, Jiangran Lyu, Jialong Wu, Fangfu Liu

High Energy Physics (Phenomenology)

LFV in flavourful SMEFT: Dimension-Six Running versus Dimension-Eight Mixing

Md Isha Ali, Siddhartha Karmakar, N Rajeev et al. (+1 more)

We study how charged-lepton flavour violation (CLFV) generated in the $\tau$ sector propagates into the $\mu\to e$ sector within the Standard Model Effective Field Theory (SMEFT). Assuming that new physics at a scale $\Lambda\gg v$ produces only dimension-6 operators that mediate $\tau\to e$ and $\tau\to\mu$ transitions, we identify and compare the three mechanisms that induce $\mu\to e$ observables. All three originate from double insertions of the two $\tau\to\ell$ operators and scale as $C_6^{\tau e}C_6^{\tau\mu}/\Lambda^4$. (T1) Renormalization-group mixing misaligns the charged-lepton mass matrix, and the rotation to the mass basis generates dimension-6 $\mu\to e$ operators. (T2) A finite, non-logarithmic remainder, which we find to be negligible. (T3) The overall divergence renormalizes dimension-8 operators, contributing after electroweak symmetry breaking. Since T1 and T3 carry no relative power of $\Lambda$, neither can be neglected at any scale, and only their comparison reveals which mechanism, and hence which $\mu\to e$ observable, most strongly probes a given operator pair. Surveying the Warsaw-basis operators, we classify representative $(\tau\to e,\tau\to\mu)$ pairs into T1-dominated, T3-dominated, and T1–T3-comparable regimes, and explain the hierarchy through the underlying mixing chains. Translating current $\mu\to e\gamma$, $\mu\to eee$ and $\mu\to e$ conversion limits into bounds on products of $\tau\to\ell$ Wilson coefficients, we find that they exceed the direct $\tau$-decay constraints by factors of ten to several hundred. Excluding either T1 or T3 would miss these limits by orders of magnitude for some pairs, and shift them by factors of two to three in others. Interpreting $\mu\to e$ data as a probe of $\tau$ flavour violation therefore requires both the dimension-6 RG running and the dimension-8 mixing contributions.
Read the full abstract

All authors Md Isha Ali, Siddhartha Karmakar, N Rajeev, Sudhir K. Vempati

General Relativity and Quantum Cosmology

Swirling spacetimes in higher dimensions: Vacuum backgrounds and black holes

José Barrientos, Adolfo Cisterna, Carlos Herdeiro et al. (+1 more)

We construct higher-dimensional swirling spacetimes in vacuum General Relativity using the hidden symmetries of the dimensionally reduced theory. Working directly with the coset-space formulation, we perform in five dimensions a reduction along the two independent axial Killing vectors, obtaining an $SL(3,\mathbb{R})/SO(3)$ sigma model and a genuinely five-dimensional background with two independent swirling parameters. The same transformation yields doubly swirling Schwarzschild–Tangherlini and Myers–Perry geometries. A distinctive feature of the new background is its multidirectional asymptotic structure. Correlated radial-angular limits reveal directional curvature singularities that are invisible in the usual large-radius limit at fixed angle. We further exhibit causal geodesics that reach such singular asymptotic ends in finite affine parameter. Finally, we describe the maximally axial extension of this construction to higher dimensions: five dimensions are maximal for an entirely axial reduction to three dimensions, while six dimensions admit an additional nontrivial translational deformation, with further dimensions contributing only flat spectator directions.
Read the full abstract

All authors José Barrientos, Adolfo Cisterna, Carlos Herdeiro, Konstantinos Pallikaris

Graphics

FeatureZ: A General Framework for Feature-Preserving Compression via Pointwise Bounds and Star Classification

Nathaniel Gorski, Xin Liang, Hanqi Guo et al. (+1 more)

Geometric and topological features, such as isosurfaces, quantiles, merge trees, and Morse-Smale complexes, are central to the analysis and visualization of scientific data across diverse domains, including medical imaging, climate science, materials science, and astronomy. However, most lossy compressors for scientific data provide only pointwise error guarantees and do not preserve derived features. Existing feature-preserving compressors are often difficult to develop and typically tailored to a single feature. In this paper, we introduce FeatureZ, a lossy compression framework for structured volumetric scalar fields that can preserve a wide class of geometric and topological features. In particular, FeatureZ frames feature-preserving compression as preserving pointwise upper and lower bounds together with a consistent classification of each point's star (i.e., its incident cells) in the underlying structured mesh. Although not all notions of feature preservation fit this framework, many features and topological descriptors targeted by existing compressors do. FeatureZ provides an efficient implementation of feature-preserving compression under this formulation. In particular, FeatureZ operates as an augmentation layer that refines the output of an existing lossy compressor. It first applies quantization to enforce pointwise bounds, and then employs an iterative procedure to ensure consistent star classification. We demonstrate that FeatureZ preserves diverse features during compression with minimal overhead, achieving compression ratios comparable to or better than methods specialized for individual features.
Read the full abstract

All authors Nathaniel Gorski, Xin Liang, Hanqi Guo, Bei Wang

Machine Learning

Bilevel optimization for data-driven learning of Koopman embeddings using kernel-based autoencoders

Joel-Pascal Ntwali N'konzi, Feliks Nüske and Stefan Klus

Koopman operator theory provides a linear framework for analyzing nonlinear dynamical systems and has become a major tool for data-driven modeling. A central challenge, however, is that finite-dimensional approximations computed by methods such as extended dynamic mode decomposition (EDMD) require the dictionary to be specified a priori. Recent machine-learning approaches address this limitation by learning the dictionary from data, predominantly using artificial neural network (ANN) autoencoder architectures. Although kernel methods offer an alternative with greater interpretability and tractability for theoretical analysis, they have received little attention in this setting. We introduce extended dynamic mode decomposition with kernel-based dictionary learning (EDMD-kDL), a kernel-based method for learning finite-dimensional Koopman embeddings directly from data. The method combines ideas from collocation methods and bilevel optimization to simultaneously learn a kernel dictionary and the corresponding Koopman approximation. We evaluate EDMD-kDL against state-of-the-art ANN-based approaches on a range of numerical experiments, including global sea-surface-temperature forecasting and learning directly from video data. Across all tested settings, EDMD-kDL achieves performance comparable to or better than the ANN-based methods. Moreover, in contrast to standard kernel methods, the proposed approach is scalable to large datasets by design since the size of the required kernel matrices depends on the number of collocation points rather than the size of the training dataset.
Read the full abstract

Robotics

Embodied Turing Machines: Stateful Code for Robot Recursive Self-Improvement

Kairui Hu, Siyuan Hu, Fangzhou Hong et al. (+2 more)

Most robot policies keep a model in the control loop: a VLA maps observations to actions, and an Agent Harness, such as Agent-as-Policy or Harness VLA queries a VLM for decision making at run time. We propose a different view: the embodied world is an Embodied Turing Machine, whose tape is the robot and environment state and rules are the policy. If this state can be represented accurately, the decision making can be written entirely in code. We therefore propose Code-Only-as-Policy (COAP): code measures and tracks the robot, environment, and task state from camera images and proprioception, and makes every decision from it. The same code applies across episodes, and different tasks share one library without a VLM or VLA in the loop. Compared with VLAs and Agent Harnesses, we analyze three advantages of COAP: (i) Explicit State: the state can be stored in code; (ii) Execution: code makes decision making controllable, recovers from failures flexibly, and runs fast and cheaply online; (iii) Extensibility: new tasks reuse, inherit, or extend the shared library, so capabilities can accumulate over tasks. These advantages make COAP a suitable medium for recursive self-improvement (RSI): coding agents develop the library in a closed loop, and each change is explicit and controllable. On RoboDojo's 42 bimanual tasks, the resulting library reaches a success rate of 70.24% without a model at test time. The upper bound of COAP lies in how accurately the state is represented for decision making and how robust the code logic is. We thus propose COAP as a new paradigm for embodied tasks; since it applies across episodes, it can also serve as an efficient data engine for VLAs and Agent Harnesses.
Read the full abstract

All authors Kairui Hu, Siyuan Hu, Fangzhou Hong, Zhaoxi Chen, Ziwei Liu

Robotics

LiteNWM: Efficient Latent World Models for Onboard Visual Navigation in the Wild

Linkai Liu, Yuntian Zhang, Zhenshan Bing et al. (+5 more)

Direct visual navigation policies generate trajectories efficiently but do not explicitly evaluate their future consequences. Generative navigation world models provide this foresight through visual rollouts, which are costly when evaluating multiple candidates. We present LiteNWM, a latent navigation world model that shares visual encoding across candidates and jointly predicts their action-conditioned future representations at multiple horizons, while a learned scorer uses these predictions to select trajectories. In offline evaluations on RECON, SCAND, and SACSoN, LiteNWM reduces macro-averaged trajectory error by 17.56% relative to NoMaD+NWM-XL and achieves a 128.00-fold end-to-end speedup on an RTX 5090. The same evaluator transfers from NoMaD to MBRA without proposer-specific retraining, reducing MBRA's macro-averaged trajectory error by 16.2%. In real-robot experiments in unseen indoor and outdoor environments, LiteNWM improves navigation success from 43.3% to 83.3% relative to NoMaD. These results demonstrate that LiteNWM can be deployed for future-aware planning and closed-loop navigation on a physical robot.
Read the full abstract

All authors Linkai Liu, Yuntian Zhang, Zhenshan Bing, Chen Chen, Lingjuan Lyu, Shangguang Wang, Mengwei Xu, Dongqi Cai