Papers

We study a rank-one spiked complex Ginibre ensemble A_N = X_N + sigma u u^*, where X_N is an N by N complex Ginibre matrix with independent entries distributed as CN(0,1/N), u is a deterministic unit vector, and sigma is a nonnegative spike strength. We focus on the maximum real part of the eigenvalues and its behavior near the non-Hermitian BBP transition. For every fixed subcritical spike strength sigma < 1, we show that the maximum real part has the same Gumbel limit as in the unspiked ensemble, after the same centering and scaling. We also derive an explicit convergence-rate correction of order 1/log N for the unspiked Gumbel law. At the critical point sigma = 1, we establish an upper bound for the rightmost eigenvalue at the N^{-1/4} scale using isotropic resolvent estimates and a net argument.

A negatively charged lead-vacancy (PbV-) center in diamond is a novel quantum system which can be operated at a high temperature above 4 K owing to its large ground state splitting. To unlock the quantum properties of the PbV- center, the enhancement of the fluorescence intensity is a key issue. Here, we demonstrate enhanced emission from PbV- centers by a factor of approximately 10 using solid immersion lenses (SILs) fabricated on rough diamond surface caused by high-temperature anneal over 2000{\deg}C. Resonant excitation reveals the narrow emission close to the transform-limited linewidth. Furthermore, we demonstrate two-photon interference using one single PbV- center in an SIL. This enhanced emission and indistinguishability will lead to further development of the PbV- center towards quantum network node applications.

Given a family $\mathcal T = (T_\lambda)_{\lambda\in\Lambda}$ of bounded linear operators acting on the same $F$-space $X$ and indexed by a set of parameters $\Lambda \subset \mathbb{R}^d$, $d\geq1$, we explore the existence and non-existence of vectors that simultaneously satisfy, for all elements of $\mathcal T$, one of the following three dynamical properties: hypercyclicity, upper frequent hypercyclicity, and frequent hypercyclicity. Our approach relies on associating to $\mathcal T$ a "Lipschitz constant function” $F$, whose growth encodes the interaction between the operators and the parameter set $\Lambda$. We show that the asymptotic behavior of $F$ determines sharp thresholds between existence and non-existence of common vectors. In the hypercyclicity setting, we identify the divergence of $\sum 1/F(n)^s$, where $s$ is the Hausdorff dimension of $\Lambda$, as a natural condition for existence, and prove complementary non-existence results, which are optimal for weighted shifts. For upper frequent hypercyclicity, we obtain criteria of non-existence in terms of series of the form $\sum 1/F(\gamma^n)$ for parameter sets in $d$-dimensions. Considered in the one-dimensional case, these criteria are optimal and allow us to contrast known existence results of Mestiri for logarithmic-type growth with non-existence beyond this scale. For frequent hypercyclicity, we establish a dichotomy showing that common vectors exist essentially only when $F$ is bounded, while any unbounded growth prevents their existence. Our results apply to a broad class of parameter sets, including self-similar fractals, providing a unified perspective on how growth conditions and geometric features of $\Lambda$ determine common dynamical behavior. Finally, in the specific context of weighted shifts, we prove that two operators do not necessarily share the same frequently hypercyclic vectors if the sequence of their weight products ratios admits two distinct non-zero cluster points, establishing the optimality of a result of Grivaux, Matheron and Menet.

We study the monoid $\mathcal W$ of based closed walks on the $1$-skeleton of the three-dimensional cube $Q_3$. The fundamental group of this graph is the free group $F_5$. We prove that every additive one-step reading, given by a sum of edge weights in an abelian group, factors through the directed transition counts, and we exhibit two closed walks with equal transition counts and different reduced loop words. Hence commutative aggregation does not determine the reduced loop word. The reduced loop word gives a proper noncommutative recognition quotient. We also determine the shortest closed walk with trivial abelianization but nontrivial degree-two commutator information. Its minimum length in the cube edge metric is $14$. We also obtain a separation between finite and unbounded memory. A two-state reading separates an order pair, while for every $k\geq1$, there is an explicit factorial pair which no $k$-state reading separates. An unbounded stack recovers the reduced loop word on every walk. For integer-valued functions on the vertex set, potential readings vanish on closed walks, while occupation binding is not rectangular. More generally, every occupation-based constraint is rectangular on a class of histories if and only if the occupation vector is constant on that class. Finally, a declared quarter-turn quaternion clock gives a second proper order-sensitive congruence, incomparable with the reduced-word quotient. The obstruction to commutative recovery, and the minimum $14$, remain valid on every hypercube $Q_n$, $n\geq3$.

In this paper, we study the Cauchy problem for the nonlinear Schrödinger equation with a nonlinearity combining a logarithmic term and a local nonlinearity which has polynomial growth. Our main result is the construction of solutions in the energy space $W_1(R^d)$ in the subcritical case. We also discuss a sufficient condition for global existence in the $L^2$-critical case in each of these spaces. Moreover, we establish a more general Cauchy theory in the space $\Sigma_\alpha$, as introduced without details in [6].

Research on 4D millimeter-wave radar perception algorithms has flourished in recent years, extending from signal processing and object detection to semantic segmentation, motion estimation, occupancy prediction, and dynamic scene reconstruction. This review organizes the field according to the evolution of perception tasks and algorithms. It first introduces radar fundamentals, data representations, and quality-enhancement methods, and then reviews object-level perception, motion and localization, local and dense spatial perception, and dynamic scene understanding. Across these directions, we compare radar-only learning, multimodal fusion, and cross-modal supervision and knowledge distillation. Particular attention is paid to how elevation, Doppler measurements, and radar physical priors are exploited across tasks. We further summarize the task coverage, input data, annotations, and evaluation protocols of existing datasets, clarifying the empirical support for different research directions. Finally, we discuss the common challenges and future directions of 4D radar perception for autonomous driving. This review provides a task-oriented perspective on the transition from sparse object perception to dynamic spatial understanding.

Traditional codecs treat every region of a frame alike; a generative layer can instead degrade the regions a viewer attends to least and reconstruct them at the client. We present PRESLEY, which extends the prior conference work ELVIS by replacing destructive block removal with adaptive in-place degradation under a removability mask, signaling per-block strength in a bit-packed side channel, and restoring via generative backbones conditioned on transmitted visual priors rather than unconditioned in-painting. We separate the problem into three goals: choosing which blocks to degrade, degrading them so the encoder spends fewer bits, and restoring them. Against its predecessor at matched rate, PRESLEY achieves a decisive mean -56.4% BD-rate reduction on delivered background quality across 13 rate ladders spanning multiple codecs and dataset families. Against pristine baselines, PRESLEY defines the operating regime of generative transport: delivering substantial bitrate savings (up to -29.4% BD-rate) and superior background quality (17/23 sequences) in the target bit-starved regime, while maintaining foreground fidelity bit-exact. We further map where the theoretical headroom in this class of architecture lies. Using an exact leave-one-superblock-out combinatorial oracle as an additive empirical bound, we show that existing complexity heuristics already capture 83.3% of bit-cost savings, bounding remaining cost-axis headroom at about 5% of total bitrate. We then identify and model the primary unaddressed axis – post-restoration damage – which disperses widely (4.9-8.4 dB). We prove that this damage is predictable before transmission (held-out rho = +0.400), establishing the feasibility of transmit-time restorability modeling and defining the roadmap for joint rate-distortion-restoration selection rules.

Socolar and Taylor asked for a single, simply connected three-dimensional prototile that forces nonperiodicity by shape alone, admitting no weakly nonperiodic tiling; the Schmitt-Conway-Danzer biprism and the three-dimensional Socolar-Taylor tile admit screw motions or a periodic stacking direction. We exhibit a rational polyhedral $3$-ball $Q$, which we call Chair44 (R44): a seven-cube chair whose $24$ exposed unit panels carry tiny square-pyramid features, and prove, as a proof submission, that $Q$ admits tilings of $\mathbb{R}^3$ by congruent copies, reflections allowed, and that every such tiling has no translational period and a symmetry group of order at most $24$; every tiling is homochiral and carries a unique infinite hierarchy of nested supertiles. The solid was designed to a reading of the aperiodic-monotile phenomenon reached with the Six Birds emergence calculus (Section 3.3), and the construction turns on a single finite test, checked by machine: the tile's own contact rule survives coarsening, so that the decoded parent tiling obeys the tile's rule and no other. The proof combines a written geometric argument, that the features force every tiling onto a registered lattice, with exhaustive finite enumerations; the companion census is replayed by two independent implementations, every finite gate is kernel-checked in Lean 4 (modulo a named compiler hook per native_decide theorem), and the written geometric lemmas and the logical assembly are Lean theorems as well, so that the theorem is kernel-checked modulo the named compiler hooks; the written proofs remain as exposition.

In this paper, we introduce layer-wise curriculum learning for efficient LLM compression. The proposed method facilitates the knowledge transfer from the teacher model to the student model, utilizing a curriculum learning approach that begins with easier optimization tasks and progressively tackles harder ones. In order to adopt the layer-wise learning in LLM compression, we partition the whole model into multiple segments consisting of layers, thereby enabling more computationally efficient knowledge transfer for LLMs. Based on our theoretical analysis of cumulative error phenomenon, layer-wise curriculum learning accelerates convergence while stabilizing the knowledge transfer process. In addition, we present a feature caching method with a multi-threading strategy to efficiently address feature misalignment across layers, maximizing GPU utilization. Consequently, our method exhibits advanced model compression performance, as well as high computational efficiency in terms of minimized memory usage and short training hours. Experiments on multiple datasets show that the proposed method achieves state-of-the-art performance while reducing GPU memory usage and training hours by more than 50% on BERT and GPT-2. Moreover, it outperforms the other pruning methods on LLaMA-family and Qwen models under the same training hours, with a lower GPU memory footprint.

Systematic generalization, the ability to solve novel problems by recombining known atomic elements, is central to human intelligence but difficult to study rigorously under controlled settings. Existing studies therefore rely on simplifications such as elemental composition, productivity-based tests, and action-explicit goals, which make systematic generalization easier to study but omit some essential aspects of this capability. To characterize what these simplifications miss, we adopt a reasoning-centered lens and introduce TranSGrid, a testbed that brings deductive, inductive, and abductive reasoning together within a unified task. Experiments with seven Transformer models on 4,800 TranSGrid instances show that all models perform much worse on TranSGrid than on a held-out test set: the largest model solves 79.6% of the test set, but only 55.3% of TranSGrid and 15.8% of the hardest subset. The gap remains within the training length range, showing that productivity alone is not sufficient to evaluate systematic generalization. Additionally, we reintroduce the other two simplifications into TranSGrid: one variant limits interactions among action effects to approximate elemental composition (reducing the inductive demand); the other makes goals action-explicit (reducing the abductive one). In both, solve rates return to roughly the test set level, showing that either simplification alone is enough to reduce TranSGrid to an ordinary held-out test set. Together, our results show that existing tasks reduce either or both of the inductive and abductive demands of systematic generalization, and that comprehensively measuring this capability requires a task that involves all three forms of reasoning.

Quantifying the heterogeneity is an important issue in meta-analysis. The conventional heterogeneity statistic, $I^2$, heavily depends on study sample sizes and tends to approach one as the sample sizes increase. To overcome this limitation, a sample size independent measure, ${\rm ICC}_{\rm MA}$, has recently been developed to quantify the heterogeneity, yet this method is restricted to meta-analysis with continuous outcomes. In practice, however, binary outcomes are more widely encountered, with the log odds ratio (lnOR) frequently adopted as an effect size to compare the effects of two treatments in terms of odds. To fill the gap, we propose a new heterogeneity measure for meta-analysis with binary outcomes by introducing a latent continuous variable model. Under this framework, the proposed measure admits a closed-form expression within the random-effects meta-analysis model with lnOR as the effect size. We define this sample size independent measure as ${\rm ICC}_{\rm MA}^{\rm OR}$, and moreover propose a new estimator, $I^2_A$, to evaluate this measure. Simulation studies demonstrate that our new estimator is nearly unbiased and aligns closely with its counterpart applied for standardized mean differences when the continuous variable is observable. A real data application further demonstrates that the proposed heterogeneity statistic for binary outcomes and its counterpart for continuous outcomes provide interpretable and coherent estimates across different outcome types.

Graph neural networks are widely used for transductive node classification, with accuracy typically measured on randomly drawn train/validation/test splits. Reported accuracy has been shown to shift substantially across different random splits of the same dataset, making published comparisons between architectures unreliable. The classical remedy in non-graph settings is stratified $k$-fold cross-validation, which ensures each test fold reflects the full class distribution of the dataset. We argue that class stratification alone is insufficient for graphs: nodes are not isolated but connected, and folds that differ in their distribution of local neighbourhood homophily expose the model to systematically different relational conditions that directly affect message-passing behaviour. The resulting cross-fold variation reflects the homophily composition of each split, inflating reported variance beyond what model behaviour alone would produce. To address this, we propose \hp{}, a topology-aware stratification procedure that treats node homophily as the primary stratification axis, aligning folds with respect to local relational consistency alongside the class marginal that standard stratification already controls. Stratifying on homophily alone does not guarantee class balance, so \hp{} incorporates class label as a secondary axis, preserving class representativeness as a natural consequence of the procedure. We evaluate \hp{} on a broad benchmark suite comprising 15 node-classification datasets spanning the full homophily spectrum and 7 GNN architectures. \hp{} achieves a mean stability rank of 1.49 compared to 2.31 for random $k$-fold, achieving the lowest mean stability rank on 13 of 15 datasets while preserving class balance close to class-stratified splits and substantially better than random. We argue that homophily-aware split construction merits broader adoption for GNN evaluation.

Generative query suggestion aims to enhance user engagement by anticipating user intents and recommending relevant follow-up queries. A central challenge is to generate slates whose individual queries are useful while the slate covers distinct intents. We propose an Intent-Driven Query Suggestion Framework with dual-stage optimization. First, intent-aware diversity modeling constructs intent-aligned supervised fine-tuning (SFT) data and uses an Intent-Aware Diversity Reward to optimize intent coverage. Second, query-level credit assignment routes individual quality signals to the corresponding query tokens while sharing a slate-level diversity signal across the slate. Experiments on a large-scale production dataset, including online A/B testing and offline evaluation, show improvements in click-through rate, query quality, and intent coverage.

In multi-robot missions such as post-disaster search and rescue, a short communication range fragments a swarm of Unmanned Aerial Vehicles (UAVs) into transient information islands. Under such intermittent connectivity, the prevailing "allocate-then-execute" paradigm–which requires global consensus before any physical movement–breaks down. This paper proposes the Communication-Constrained Online Performance Impact (CC-OPI) algorithm, an event-driven method that interleaves task negotiation with physical execution. CC-OPI replans only at discrete physical and topological events and integrates two further elements. The first is a pair of cost-evaluation metrics adapted to dynamic topologies–one with a spatial locality penalty that promotes regionalized operation, the other with a deadline-aware urgency term–complemented by a non-preemptive state lock that shields each UAV's ongoing action. The second is a decentralized fault-tolerance layer that pairs version-based state synchronization with a global-time-driven emergency pool. We establish that CC-OPI terminates in finite time, free of stale-completion deadlock and of unbounded reassignment within the mission horizon. In simulations at a 250 m communication radius, CC-OPI sustains a task completion rate of about 0.80: it leads a matched online execution of the unmodified Performance Impact (PI) and Consensus-Based Bundle Algorithm (CBBA) rules by about seven percentage points, exceeds the naively transferred static baselines by roughly 20 points, and remains within several points of PI and CBBA under full connectivity. Within the tested settings, CC-OPI degrades gracefully as connectivity weakens and absorbs packet loss, terrain occlusion, and runtime task arrival. The price is more messages and some redundant travel–a deliberate trade-off of efficiency for robustness.

Autoregressive transformer decoding is constrained by irregular key-value (KV) cache movement on tiled accelerators. Prior compression and DRAM-placement systems still concentrate traffic on centralized memory paths that bottleneck long-context serving. We present MeshKV, a KV cache fabric that moves blocks as packetized flows over a lightweight NoC. It co-designs (i) TaKV affine striping to spread homes and cut hotspot load, (ii) Mare multicast with verified duplicate suppression, and (iii) Pad, which overlaps prefetch, tile multiply, and streaming softmax behind credit-aligned FIFOs. Together they convert bisection back-pressure into useful KV transfer. On our 8x8 FPGA implementation with LLaMA-2-7B and Mistral-7B at 8K-32K, MeshKV reduces interconnect traffic by up to 58%, improves KV bandwidth utilization by 2.1x, and delivers up to 1.9x multi-stream throughput.

In-storage computing (ISC) reduces host–storage data movement by executing computation inside computational storage devices (CSDs). For multi-stage applications, realizing these benefits requires coordinating data placement, I/O–compute overlap, and device-resident state across the workflow, yet existing interfaces lack a unified abstraction for these decisions. We present Epic, an NVMe-based ISC stack that provides this abstraction by capturing data residency and lifetime in the program: location types declare logical residency, dataflow derives lifetimes for intermediate values and operation state within an invocation, and a keep primitive extends selected state across invocations. These semantics expose the complete offloaded workflow as a located, stateful dataflow. A storage-aware compiler transforms this workflow, performs movement-aware logical mapping and fusion, and exposes I/O–compute overlap; a runtime completes the plan using execution-time information, asynchronously binding work to physical resources and managing device-resident state. Across 12 file-scanning, database, and machine learning workloads, Epic is 1.6$\times$ faster on average than the strongest of five prior ISC systems, while achieving 4.2$\times$ speedup on average and up to 16.1$\times$ over the corresponding host baselines, and reducing application-side code by up to 14$\times$ in our implementations.

The present paper studies the initial-boundary value problem of the nonhomogeneous incompressible two-phase magnetohydrodynamic model with Landau potential in a bounded smooth domain in $\mathbb{R}^d$($d = 2, 3$). More precisely, we construct the existence of local in time in three dimension and the global existence of strong solution in two dimension with arbitrary large data and bounded density. In addition, the uniqueness of the solution is proved through the weighted energy estimates, the shift of integrability method and Lagrangian approach.

Robotic systems operating in dynamic environments require visual perception that evolves continuously with the incoming sensory stream. Event cameras provide microsecond temporal resolution and asynchronous sensing, but most learning-based methods accumulate events into frames or temporal bins, introducing an integration delay that can limit fast reaction. Here we propose REACT, a fully spiking state-space model for event-driven temporal perception that processes raw events one by one, without temporal accumulation. REACT uses a complex-valued spiking neuron, C-SiLIF, whose continuous-time dynamics are driven by the physical inter-event interval, allowing its internal state to evolve at the temporal resolution of individual events. We evaluate REACT on gesture recognition and time-to-collision (TTC) estimation from full-field event streams, without a target bounding box or localization input. On EvTTC, REACT achieves a 9.59% relative TTC error with 4.6 ms end-to-end inference latency, within 0.15 percentage points of the best learned method while requiring no target prior. At the dataset's mean approach speed, this latency corresponds to only 4 cm of vehicle motion, compared with 1 m for the fastest competing learned method. REACT further supports anytime TTC prediction, zero-shot transfer to a different driving sequence, and INT8 quantization, reducing the estimated energy consumption from 18.5 to 2.8 mJ per 32,768 events. These results show that event-driven spiking state-space dynamics can provide low-latency, continuously updated temporal perception for reactive robotic systems.

AI applications have shifted from single, monolithic foundation models (FM) to compound agentic systems. Yet today's stacks remain fragmented: even as protocols (e.g., MCP, A2A) ease tool/agent connectivity, each framework embeds an implicit runtime for state, memory, budgets, and guardrails, making behavior non-portable and governance brittle. It mirrors computing before operating systems, when every program re-implemented basic services. This position paper argues that the field now needs a Foundation Model Operating System (FMOS) – a system layer that virtualizes FM interactions analogous to how virtual machines abstract physical hardware, giving applications the illusion of dedicated, trustworthy FM instances with effectively unbounded capabilities. Internally, the FMOS orchestrates knowledge across memory tiers, model selection and resource allocation, and verification and policy enforcement. Like the human brain switching between fast intuition and slow deliberation, the FMOS learns when to intervene and when to let inference proceed directly and continuously adapting its policies based on operational experience.

We study federated soft clustering over federated learning (FL) networks of devices that each hold a private local dataset and fit a personalized Gaussian mixture model (GMM). Generalized total variation minimization (GTVMin) couples the local maximum likelihood problems through a graph regularizer that penalizes a discrepancy between the models of connected nodes. The choice of discrepancy measure is a key design decision: we compare a squared Euclidean distance between model parameters, which requires component matching, with two measures that compare the local model distributions directly and hence need no matching: a Monte-Carlo approximated Kullback-Leibler (KL) divergence and a closed-form maximum mean discrepancy (MMD). All three resulting GTVMin instances are optimized by synchronous projected gradient updates; for the smooth MMD instance we provide a convergence guarantee to stationary points. We characterize their computational cost and evaluate their robustness to data heterogeneity.

Existing Security-oriented Video Understanding (SVU) systems assume a closed world, \ie static category sets, abundant labels, and the premise that all event types are known upfront. Real-world security-critical events break these assumptions: they follow long-tailed distributions, new types emerge continuously, and critical security events may offer only a few samples. We formalize this gap as \textbf{Lifelong Evolving Task for Long-Tailed Security-Critical Events in Videos ({\boldmath$L^2$}-SCE)}, a new task that requires VLMs to continually classify and temporally localize newly emerging security-critical events from scarce samples without forgetting previously learned events. Furthermore, \task reveals two critical challenges: (1)~Intra-Event Scarcity, where extreme data scarcity may weaken both classification and temporal localization for new events, and (2)~Inter-Event Interference, where cross-event feature entanglement and representation drift may strengthen catastrophic forgetting. On this basis, we propose \method, a causal-enhanced approach orchestrated end-to-end by an Agentic Controller with self-reflective closed-loop control, which includes two core modules: the Intra-Event Causal Video Generation module (CVG) and the Inter-Event Causal Decoupling and Alignment module (CDA), to address the above two challenges, respectively. Especially, this paper constructs a \task dataset to simulate real-world incremental conditions. Extensive experiments on our benchmark demonstrate the advantages of \method over several advanced baselines. These justify the importance of the proposed \task and the effectiveness of \method in classifying and temporally localizing emerging security-critical events from scarce samples.

Underwater visuomotor policy learning has focused primarily on single manipulators, while bimanual imitation learning has been studied largely in air. We present ULOHA, an underwater bimanual robot learning platform that combines custom-designed leader–follower hardware with software extensions to LeRobot, integrating teleoperation, multi-view sensing, demonstration collection, policy training, and autonomous deployment. Real-robot experiments demonstrate a range of coordinated underwater bimanual behaviors, including inter-arm transfer, shared-object manipulation, and buoyancy-driven interception. We evaluate ACT, Diffusion Policy, and the vision–language–action model SmolVLA on the platform. We investigate how learning methods and execution strategies developed for manipulation in air perform underwater, examining bubble disturbances, buoyancy-driven object motion, action-execution horizons, and real-time chunking. A separate single-arm study examines policy transfer between air and water and shows that demonstrations spanning both media support execution in both under the tested conditions. ULOHA provides a unified experimental platform for studying underwater bimanual robot learning under the coupled perceptual and physical effects of underwater environments. Additional material: https://mertcookimg.github.io/uloha/

Large Language Models (LLMs) are increasingly adopted for compliance and legal reasoning tasks, yet their outputs often lack explicit grounding in legal logic and evidence. We present Code-as-Auditor, an LLM-based framework that extends the model's reasoning capability toward structured and evidence-grounded compliance assessment. The framework translates regulatory information into (1) formalized checklists and executable decision trees, encoding regulations and conditions as interpretable code structures. During inference, each checklist item is (2) dynamically expanded into factual and counterfactual questions, guiding the model to reason over case-specific evidence and potential violations. This process establishes a reasoning pipeline that proceeds from evidence identification, through rule application, to final decision-making, while a self-verification loop improves the logical consistency of the generated code and the traceability of outcomes. Experiments on privacy and data protection scenarios demonstrate that Code-as-Auditor delivers more accurate and evidence-backed evaluations, enabling automated compliance regulation checking grounded in explicit regulatory criteria.

In this work, we propose a notion called effective distance to characterize the essential gravitational effects on quantum coherence of neutrinos. Employing effective distances, we prove that one can always map neutrino-oscillation phases in asymptotically flat spacetimes onto the phases in flat spacetime, which is considered as a nontrivial realization of equivalence principle. Hence, the maximal amount of quantum coherence, quantified by the Leggett-Garg (LG) parameter is independent from background spacetimes, namely, the paradox that the gravitational effects on quantum coherence of neutrinos are measure-dependent (Ettefaghi et al.[1]) does not arise. Following the same route, we preclude the gravitational amplifying or damping of amplitudes of the quantum information quantities for special masses of black holes (Wang et al. [2]). To demonstrate furthermore the independence of quantum coherence from background spacetimes, we examine the proposal using the LG parameter to identify quantum-corrected models of spacetimes. For different emission positions of neutrinos and various quantum parameters, we find no consistent patterns in oscillating LG curves to discriminate a quantum metric from the classical one.

Let $\psi$ be a real Arthur parameter and let $\psi_2$ be the unipotent parameter in a fixed Jordan decomposition. Adams, Ionov, Mason-Brown, and Vogan proved that the microlocal packet of $\psi$ is contained in the support of the two-step Jordan induction of the packet of $\psi_2$, and conjectured equality. We prove the reverse inclusion. The argument first passes to a sufficiently positive translate, where the relevant connected components of the ABV spaces have a common flag-variety model. Perverse devissage reduces the problem to a singular-support implication, which follows from the incidence calculation underlying the AIMV microstalk comparison. Coherent continuation for individual packet members then returns the result to the original parameter.

Collecting data for manipulation with high-DOF hands is challenging, as interfaces must capture rich hand motion while rendering the contact interactions essential for precise manipulation. Existing data collection approaches face a trade-off: teleoperation ensures deployment consistency but lacks force feedback, while handheld (in-the-wild) systems provide natural force transparency but introduce a visual embodiment gap at deployment. We present DITTO, a Dexterous Interface for Transparent TeleOperation, which resolves this through the anatomically informed co-design of a dexterous 7-DOF robotic hand and a kinematically equivalent motorized exoskeleton. A 1-to-1 actuator mapping between the exoskeleton and robotic hand enables handheld (in-the-wild) data collection and bilateral teleoperation with joint-level force feedback unified in a single platform. We demonstrate that the DITTO exoskeleton spans the operator's natural index-to-thumb workspace, and showcase DITTO's dexterous capabilities via learned policies on contact-rich tasks.

Two dominant tokenisation algorithms are used by modern language models: byte-pair encoding (BPE) and UnigramLM. These differ along two orthogonal axes: their optimisation objective (compression vs. log-likelihood) and their search procedure (bottom-up merging vs. top-down pruning). Existing comparisons confound these axes, making it unclear whether their observed differences stem from what is being optimised vs. how it is being optimised. We disentangle the two by introducing two new tokenisation algorithms that complete this 2x2 design space: BottomUpLL, a bottom-up likelihood-based tokeniser, and TopDownComp, a top-down compression-based tokeniser. We train language models with tokenisers produced by each algorithm, varying: model size, vocabulary sizes, and domain (English-only vs. multilingual). Evaluating models on bits-per-byte, we find that the search procedure – not the objective – is the dominant factor: bottom-up tokenisers consistently achieve lower bits-per-byte in most settings. Evaluating models on the BLiMP task, however, shows no consistent relationship between design choice and performance. Overall, our results disentangle the effect of tokeniser design choices on language modelling performance, offering concrete guidance for their more principled construction.

Direct preference alignment methods are widely used to align large language models (LLMs) with human preferences because of their computational and memory efficiency. However, likelihood displacement motivates alternative ways to extract information from preference pairs with small likelihood margins. In this paper, we propose and analyze Comparison-based Preference Optimization (ComPO), a zeroth-order alignment method based on comparison oracles. ComPO extracts directional information from these pairs without directly optimizing a differentiable preference loss on them. We establish a convergence guarantee for its basic offline scheme under smoothness, gradient sparsity, and compatibility between the oracle and a latent objective. We further introduce online ComPO, which retains the offline comparison mechanism and uses unlabeled policy generations for reverse-KL control relative to a reference policy. Following the coverage perspective of preference fine-tuning, we establish a performance guarantee for a basic constrained scheme under local coverage and in-distribution pairwise reward accuracy. Experiments on Mistral, Llama, Gemma-2, Qwen3, and Gemma-3 models demonstrate improvements over existing direct alignment methods, including length-controlled win rates, with pair-level diagnostics providing evidence consistent with mitigating likelihood displacement.

Intelligent systems that act in the world require image understanding that is both comprehensive and spatially grounded. Current vision-language models (VLMs) can generate fluent and detailed image captions, but reliably associating them with image pixels remains challenging. Existing methods that combine dense captioning with pixel-level grounding often produce either incomplete descriptions or inaccurate segmentation masks. We study this problem through panoptic grounded captioning, a task that requires a VLM to describe both foreground objects and background regions while grounding each referring phrase with pixel-level masks. We make three contributions. First, we introduce PanoCaps, a human-annotated benchmark constructed from panoptic segmentation datasets. It provides dense captions with near-complete pixel coverage and image-text alignments at the entity level, supporting both training and evaluation. We further propose a phrase-mask matching protocol and a generalized Panoptic Quality (gPQ) metric that jointly evaluates textual and mask agreement. Second, we formulate phrase grounding as selection from a phrase-conditioned pool of mask proposals and introduce PANORAMA, a VLM that conditions a pretrained segmenter on contextualized phrase representations to obtain candidate masks and learns to select those corresponding to each phrase. Training this interface jointly with caption generation enables PANORAMA to produce high-quality masks while allowing each phrase to refer to a single region or multiple instances. Third, PANORAMA achieves the best overall grounding on PanoCaps and matches or exceeds specialized models across several pixel-level grounding tasks. Experiments show that our method produces precise entity-level segmentations while maintaining detailed, mask-consistent captions. Code, data and models are available at https://www.di.ens.fr/willow/research/panorama/.

World models endow perceptual systems with the ability to predict how scenes evolve under interaction. They are most beneficial when trained on diverse volumes of data, to instill a rich prior into downstream applications. Existing methods typically require robot action labels to learn action-conditioned 3D dynamics, which excludes web video data from the training pool. We study 3D point track completion as a pre-training objective for learning transferable 3D dynamics without robot data. Given a single RGB-D observation and sparse partial 3D trajectories (tracks), we predict future 3D tracks of all observed points. We show this objective produces a rich 3D dynamics prior, without requiring robot action labels. We contribute a diverse dataset of 2.9 million synthetic frames spanning deformable, articulated, and rigid objects, and use it to train PointZero. We show that a flexible and expressive transformer, PointZero, outperforms prior methods on the same data. We demonstrate the utility of our pre-training objective by post-training PointZero for two downstream applications: (1) action-conditioned 3D dynamics prediction and (2) imitation learning. When fine-tuned to condition on end-effector pose, PointZero outperforms the baselines on the recent PGND 3D dynamics benchmark. When fine-tuned to predict robot actions and 3D tracks, PointZero outperforms or matches the baselines on 6/7 simulated and real-world robot manipulation tasks. We furthermore evaluate training PointZero from scratch to isolate the benefits of our proposed architecture from those of our proposed pre-training objective and dataset. We release the dataset, checkpoints, and full training recipe.