Papers

Sparse longitudinal CT follow-up limits lesion-size forecasting when only a few prior observations are available. We constructed a five-visit DLT-derived same-lesion trajectory benchmark from DeepLesion and Deep Lesion Tracker (DLT), yielding 205 trajectories from 129 patients. We compared an exploratory conventional sparse-to-final analysis with a primary fixed visit-index horizon design predicting the common log change from T3 to T4 while progressively adding earlier observations, evaluating predictive accuracy, uncertainty reliability, post-hoc conformal interval calibration, subgroup performance, and Gompertz-inspired trajectory regularization. The evaluated methods showed partially overlapping point-prediction accuracy but distinct uncertainty behavior. Mean held-out RMSE across ten training seeds was 0.4726, 0.4305, 0.4499, and 0.4513 for m = 1, 2, 3, 4, indicating the lowest mean RMSE at m = 2; additional history did not improve RMSE. At m = 4, raw Cohort-Level Feature GP coverage was near the 95% nominal level, whereas MC Dropout, Deep Ensemble, and residual-scale intervals were conservative. Patient-level conformal calibration generally produced near-nominal or conservative coverage at the cost of wider intervals. Patient-grouped development cross-validation selected lambda* = 0 for the Gompertz-inspired term. A global population reference frequently opposed lesion-level change directions, and prediction difficulty varied across anatomical subgroups. Overall, additional historical observations provided limited predictive benefit once the prediction horizon was controlled, while predictive accuracy, uncertainty reliability, and trajectory consistency did not necessarily improve together, and should be evaluated jointly in sparse longitudinal imaging.

Generative AI can improve students' programming performance, but successful task completion may not reflect what they retain. We examined performance, retention, cognitive load, and ownership in a controlled between-subjects experiment with 59 undergraduate computer science students, 55 were retained for analysis. Participants completed three introductory C programming tasks with access to ChatGPT-4.5 or conventional web search without generative AI. We measured task performance, self-reported mental effort and difficulty, pupillary responses, heart rate variability, and ownership, and assessed cued recall immediately and 48 hours later. ChatGPT-assisted students achieved higher coding scores (89% vs. 69%) but lower recall scores immediately (41% vs. 53%) and after 48 hours (39% vs. 52%). There was no significant difference in the loss of recall information over 48 hours between the groups. Self-reported mental effort increased less across tasks in the ChatGPT condition (Holm-adjusted p = .047), and students attributed less of the submitted code to themselves (45% vs. 81%). Confirmatory physiological tests did not detect significant differences in trajectories between conditions; substantial data loss limits their interpretation. These findings reveal a gap between assisted task performance and subsequent recall and sense of ownership in this setting. They motivate the need for assessment practices and AI learning tools that require students to explain, retrieve, and contribute to the work they submit as active participants in their education.

Organizations deploying agentic artificial intelligence must determine more than whether a model is trustworthy; they must establish what to validate, control, and observe for a use case to deliver its intended outcome while meeting applicable obligations. This paper proposes AI-GRACE (Agentic Intelligence-Governance, Risk, Assurance, Controls, and Evidence) as a use-case operationalization framework connecting organizational governance with technical implementation. The proposal draws on professional observations and a purposive synthesis of standards and literature, using design science to frame the method contribution and situational method engineering to guide contextual tailoring and reuse. The framework establishes objectives and obligations and then assesses risks in seven proposed domains, including mission and value realization. It derives requirements for assurance before deployment, runtime controls, and evidence, which guide capability qualification, gap assessment, and a logical architecture. An Agent Operating Envelope specifies permitted actions and escalation conditions, while Risk-Aligned Independence Levels (RAIL) summarize the authorized independence. A fictional retail banking application illustrates the method. The contribution is a traceable basis for deciding what an organization must implement, what it already supports, and what remains unresolved. Empirical evaluation must establish whether it improves deployment decisions, efficiency, and reuse.

Ensuring the correctness of LLM-generated code is a core challenge for modern software engineering. Benchmarks for agentic code generation check correctness with held-out test suites, which are inherently incomplete and increasingly susceptible to memorization. Formal verification avoids both problems, but existing work covers only standalone tasks whose specifications are given as input, not real issues, which touch large repositories and state intent in vague natural language. We present Benchproofer, a pipeline that turns a coding task with a known correct patch into a formally verified one: it writes a specification for the new code, summarizes the existing functions that code calls with axioms, and admits an instance only after mechanical and adversarial gates agree. Applying it to SWE-bench Verified yields SWE-Proof, 500 real issues whose correctness is formally verified rather than tested, and it extends to SWE-bench Pro. Evaluating Claude Opus 4.8, we find that verification catches what tests miss: a quarter of test-passing patches admit counterexamples, which a structured natural-language specification does not fix, while a correct formal one lifts resolution from 85% to 95%. Writing that specification is the hard part: an agent that must write its own gains nothing over an unaided baseline, and only 56% of those specifications pass our audit. The usual failure is faithfulness, a specification that constrains part of the required behavior and leaves the rest free. Specification quality still tracks the outcome, failing on 92% of unresolved instances against 51% of resolved ones, making faithful specification synthesis a concrete open problem.

We study the asymptotic density of $\mathbb{F}_q$-point-free linear sections of geometrically irreducible projective varieties over finite fields. We then apply these results to rank-metric codes via determinantal varieties. Our approach recovers the known cases in which the density tends to $0$ or $1$ and determines the limit in the cases where it was previously unknown. To compute these previously unknown limits, we extend the notion of quasireflexivity to higher-dimensional varieties and show that determinantal varieties satisfy this property. This allows us to invoke the Chebotarev density theorem for varieties over finite fields to obtain the desired estimate.

Agent models are frequently evaluated one decision at a time, where the model predicts the next action based on the gold interaction history, which is scored against a reference. We investigate whether improvement under this protocol is predictive of improved autonomous workflow execution. We study pre-SFT and supervised fine-tuned (SFT) Qwen3 models at 4B and 14B parameters and Gemma 3 models at 4B and 12B parameters on multi-turn customer-support workflows. We find that SFT consistently improves text-turn success, and that overall next-turn success increases for every model under gold-history evaluation. However, these improvements do not transfer to autonomous workflow execution. Tool-specific gains also vary across metrics and models. None of the four SFT models succeeds under holistic workflow evaluation, with strict trajectory completion reaching at most 10.4% workflow success. Our results show that next-turn evaluation is not a reliable proxy for workflow success, motivating separate reporting of text quality, local action correctness, tool execution, and end-to-end task completion.

Accurate initialization is essential for reliable visual-inertial odometry (VIO), but it is often ill-conditioned under degenerate motions. Existing methods typically require restrictive excitation motions to ensure sufficient observability or rely on computationally expensive 3D structure reconstruction, limiting efficient and practical deployment. To address these limitations, we propose SLIM-init, a structureless monocular VIO initializer that directly exploits geometric constraints from tracked 2D line features without explicit 3D landmark reconstruction. Specifically, SLIM-init leverages line-derived vanishing points (VPs) as translation-invariant orientation cues to provide robust rotation-only constraints under degenerate scenarios such as low-parallax or translation-dominant motions. It further incorporates a line epipolar residual to constrain translation and a line-normal projection residual to improve the conditioning of linear alignment, enhancing the accuracy and robustness of initial state estimation. Extensive experiments on a public benchmark and challenging custom degenerate-motion sequences demonstrate improved accuracy and robustness over state-of-the-art initialization methods. The source code is available at: https://github.com/cjunwan/SLIM-init.

Low-friction surfaces challenge bipedal locomotion by limiting the contact forces available during stepping. Inspired by penguin waddling, we investigate how lateral torso motion and center of mass (COM) placement affect locomotion as surface friction changes. Using a five-actuator biped, we compare an upright-gait strategy with a penguin-inspired torso-over-stance-leg strategy across multiple COM placements in simulation and hardware. In the 3-D simulator MuJoCo, we sweep through sinusoidal leg and hip actuation parameters across four friction coefficients mu = 0.1, 0.3, 0.5, 0.7. In simulation, torso-over-stance-leg motion produces more successful controllers and higher forward speeds at low friction, with the highest speed occurring for the high-COM configuration. Hardware experiments show the same low-friction speed trend: at mu=0.12, torso-over-stance-leg motion increases forward speed and reduces cost of transport at both tested COM ratios, and the higher COM also improves both measures. The high-COM penguin configuration is the fastest and most energy efficient while maintaining low sideways foot motion. At mu=0.45, the COM trend reverses: the lower-COM configurations are faster and more energy efficient, while gait strategy has little effect on forward speed but still changes sideways foot motion. These results show that the effects of lateral torso motion and COM placement depend on the available friction, and that forward speed, energy use, and slip-related foot motion can be modulated with a penguin-inspired torso motion on hardware.

Audio large language models (AudioLLMs) operate reactively, responding only when queried. We introduce proactive audio assistance, where an AudioLLM monitors an audio stream and autonomously decides when to alert the user from a single natural-language intent, motivated by wearable applications for Deaf and Hard of Hearing users. We propose Interrupt and Silent Modeling (ISM), a model-agnostic paradigm that embeds proactive decisions into LLM decoding via two special tokens: <interrupt> and <silent>, capturing four states: onset detection, sustained-relevance triggering, irrelevance suppression, and de-duplication. Applied to Qwen2-Audio-7B, ISM achieves 99.6% interrupt F1 and perfect de-duplication recall on ESC-50. On noisy Epic-Sounds kitchen audio, ISM achieves the highest interrupt F1 without domain-specific training, the only method maintaining strong onset detection without over-triggering or over-suppression. Streaming evaluation confirms real-time viability with 3.5-second average latency.

The Abstraction and Reasoning Corpus and related benchmarks evaluate whether AI models can solve novel reasoning tasks, but often leave unclear whether success reflects inference of the intended underlying rule or reliance on shortcuts. We address this gap by studying test-time task embeddings in Vision ARC (VARC), a model in which a pre-trained backbone is complemented by a trainable embedding representing the transformation rule. In the original VARC, test-time training (TTT) is jointly applied to the backbone and task embedding. Here we introduce a novel two-step TTT protocol: first finetune only the task embedding (Embed-TTT), then freeze it and finetune the backbone. Across ARC-AGI-1, ConceptARC, and two controlled datasets with known rules, Embed-TTT consistently yields improved task embeddings, ones that align better with underlying task rules, improve embedding-based retrieval, and enable accurate linear probing of known rules. Qualitatively, Embed-TTT identifies more semantically meaningful relations between test and train tasks on ARC-AGI-1. We also show that optimizing only task embeddings (less than 0.01% of model parameters) already solves a non-trivial fraction of ARC-AGI-1, ConceptARC, and Mini-ARC tasks, while the full two-step pipeline improves final performance. Finally, we show that Embed-TTT recovers the underlying geometric structure of parametric rules and learns compositional capabilities that enable rule-wise interpolation, but not extrapolation. These findings support a clearer separation between rule induction and rule execution in ARC-like evaluations, motivating benchmarks that better distinguish in-distribution from out-of-distribution rules.

Real-time multichannel speech extraction (MSE) under diffuse noise conditions is an important task with a wide range of applications, such as speech recognition and hearing aids. In this paper, we propose online algorithms for independent low-rank matrix analysis (ILRMA) and rank-constrained spatial covariance matrix estimation (RCSCME). Previously, we proposed a real-time extension of the RCSCME-based method: an MSE method based on ILRMA and RCSCME using the blockwise batch algorithm. However, it assumes that the spatial characteristics are stationary within a single batch, and thus, in dynamic situations where the target speaker moves, its performance may degrade. To address this problem, we derive the online algorithms for ILRMA and RCSCME in the following three steps. First, we formulate framewise cost functions for ILRMA and RCSCME on the basis of maximum weighted likelihood estimation. Second, we derive the update rules for the framewise cost functions on the basis of auxiliary-function techniques. These naive update rules are computationally costly for real-time execution on a practical machine. Thus, we finally derive the online algorithms by approximating some intermediate parameters with their estimates. Furthermore, we propose stabilization and further acceleration techniques for these online algorithms. In experiments, we simulate situations where a target speaker is stationary or moves and show that the proposed method achieves superior speech extraction performance compared with conventional methods. In addition, using real-world recorded signals, we demonstrate the effectiveness of the proposed method in practical scenarios.

Neural morphological generation systems often achieve high aggregate accuracy on benchmark datasets, yet such performance can conceal systematic errors clustered in rare morphological subclasses. We present an orthography-aware diagnosis of Japanese past-tense verb inflection, treating hiragana not merely as a transcriptional medium but as a representational system that encodes morphophonological structure. Using two character-level Transformer architectures evaluated across five random seeds, we show that although both systems exceed 97% aggregate accuracy, a single structurally specific irregular subtype, verbs whose stems end in /e/ and require gemination before the past-tense suffix and make up fewer than 1% of the data, accounts for a disproportionate 30-43% share of residual errors and contributes roughly 34-48x its prevalence to total errors. We then move from diagnosis to causal isolation: controlled ablation experiments show that removing this subtype alone produces larger accuracy gains than removing all irregular verbs combined. These findings indicate that error concentration in neural morphological learning is not driven by irregularity per se, but by the interaction between extreme low-frequency morphological patterns and specific orthographic processes. We argue that morphological evaluation should incorporate fine-grained subclass analysis, and discuss implications for data-efficient, developmentally plausible language model pretraining.

Service applications for human-robot interaction are commonly written against the hardware-specific interfaces of one platform, so a change of hardware forces a rewrite of the application. The Robotic Interaction Service (RoIS) Framework 2.0, standardized by the Object Management Group (OMG), addresses this fragmentation by defining a platform-independent model in which Service Applications interact with Human-Robot Interaction (HRI) Engines through standardized interfaces and hardware-independent symbolic messages. A specification alone, however, does not provide the maintained implementation, Software Development Kits (SDKs), and adapters needed for practical adoption. This paper presents OpenRoIS, a community-driven open-source middleware providing a concrete implementation of the RoIS Framework 2.0. It takes the position that an openly developed, paradigm-neutral implementation is what carries the standard from specification to practice. OpenRoIS contributes a recursive engine architecture in which a single engine class realizes the main and sub HRI Engine roles, an internal five-method component contract distinct from the five external RoIS interfaces, a mapping of those interfaces onto JSON-RPC 2.0 over WebSocket, a single-source-of-truth type pipeline that generates three consistent language stacks, TypeScript and C# client SDKs that include web and Unity support, and a Python adapter SDK that includes ROS 2 support. Through the common RoIS interfaces, a Service Application can address physical robots and virtual agents over the internet. All source code, interface types, and documentation are released under the Apache-2.0 license and openly developed at https://openrois.org/.

This paper addresses the challenges of dynamic scene synthesis from sparse-view videos. Existing methods employ geometric priors, adaptive optimization, or density-control strategies to improve 4D Gaussian modeling under sparse observations. However, they cannot fundamentally resolve the ill-posed problem caused by insufficient observations and missing scene information. Moreover, sparse-view 4D Gaussian Splatting (4DGS) often suffers from poor geometric initialization: with only a few input views, COLMAP typically reconstructs sparse and incomplete point clouds, leaving large scene regions without sufficient Gaussian support and making them difficult to recover through subsequent optimization. To address these limitations, we propose a novel iterative refinement framework based on a video diffusion model to improve the completeness and consistency of dynamic 4D scenes. Specifically, we first estimate multi-view depth maps and fuse them into dense point clouds to provide more complete geometric initialization for a dynamic 4DGS representation. We then employ a pretrained video restoration model to refine sequences rendered along novel camera trajectories at different time steps. The restored sequences serve as pseudo-supervision to regularize and iteratively refine the 4DGS representation. Experiments on a widely used benchmark dataset demonstrate that our method substantially outperforms existing baselines, achieving nearly a 2 dB PSNR improvement over the previous best-performing method.

In this work, we develop a second-order, linear, decoupled, and structure-preserving numerical scheme for the modified Landau–de Gennes model of smectic-A (SmA) liquid crystals. The main contributions are threefold. First, to the best of our knowledge, we propose the first integration of the generalized scalar auxiliary variable (GSAV) approach with a second-order exponential time-differencing Runge–Kutta (ETDRK2) discretization, leading to a second-order GSAV–ETD2 scheme. Second, we prove that the proposed scheme satisfies an unconditional energy-dissipation law, thereby closing the theoretical gap in the energy-stability analysis of second-order GSAV exponential integrators of this class. Third, by deriving a coercive discrete reformulation, we establish a fully discrete error estimate without imposing any coupling condition between $\tau$ and $h$, achieving the optimal convergence rate $\mathcal{O}(\tau^2+h^2)$. Numerical experiments are presented to verify our theoretical results and to simulate the self-assembly dynamics of the SmA phase.

The actor model - state isolation, data-race freedom, and sequential single-message reasoning - has long been dismissed as unsuitable for high-frequency trading (HFT): actors seem to imply many threads, a mailbox per actor, and a heap-allocated message plus a context switch per interaction, overhead incompatible with a microsecond budget. This paper argues the dismissal is wrong for co-located actors, with a deployed, measured implementation: kaspar-hft, an open-source C++20 framework. Four extensions adapt the model for HFT: fast_send, a synchronous delivery mechanism in which the sending thread runs the receiver's handler inline and returns the reply as a value; actor groups, which co-schedule actors on one thread behind a shared mailbox; per-actor selectable mailbox queues; and a memory pool. fast_send has receiver transparency: the handler is written identically for synchronous and asynchronous delivery and does not depend on which was used or which thread runs it. A grouped synchronous chain runs on one thread, cutting scheduler context switches from O(N) to O(1); a thread-local call-chain test detects cyclic invocation on one thread before any lock is taken, while a cycle spread across threads deadlocks. Microbenchmarks put the synchronous round trip at tens of nanoseconds. On a live CME market-data feed (ES, NQ, ZN futures), when the socket-reader thread decodes each packet and updates the book itself, socket-to-book medians for book updates are 0.8-1.1 microseconds, and a fast_send hop is about 1% of that. The shared-queue group also yields a production/simulation duality: the same actor code runs unchanged in live trading and deterministic backtest.

Large language models (LLMs) are moving onto mobile devices for increasingly diverse workloads over text, images, video, and audio. These applications often require long contexts, making the Key-Value (KV) cache a dominant memory bottleneck because it grows linearly with sequence length and is accessed at every decoding step. Prior work reduces KV-cache footprint through low-rank compression, token eviction, or flash offloading, but the resulting reconstruction overhead, irreversible token loss, or I/O stalls can offset the benefit of saving memory. We present TierKV, a mobile LLM inference framework built on Predictive Multi-Tier Cache Optimization (PMCO). Before decoding starts, PMCO predicts future cache demand from prefill hidden states and jointly assigns tokens to exact, low-rank, and flash-offloaded tiers under the device memory and accuracy budgets. This formulation retains access to the full context, removes the circular dependency of reactive eviction, and admits a closed-form solver that selects tier boundaries and per-layer ranks at runtime. Across eight text, vision, and audio models on three mobile SoCs, TierKV improves prefill throughput by up to 17.6x over existing mobile LLM frameworks, reduces RAM-resident KV cache by 12.5-34%, thereby enabling substantially longer contexts under the same memory budget, while incurring only minor accuracy degradation.

Safety-critical control systems must contend with disturbances whose extreme deviations occur far more frequently than classical light-tailed models predict. Existing stochastic MPC (SMPC) formulations tighten constraints using an assumed distribution, a moment bound, or a finite scenario sample, each of which degrades under an unknown heavy-tailed disturbance. This paper develops an SMPC formulation for linear systems under heavy-tailed disturbances that is only assumed to be regularly varying, which allows an explicit extreme value theory (EVT) characterization of the tube error tail that is asymptotically exact. We further show that closed-loop dynamics induce temporal clustering of rare excursions across the prediction horizon, and characterize this clustering through a closed-form extremal index estimable from data. The resulting clustering-corrected constraint bounds the probability of a rare-event episode over the horizon, rather than only the marginal per-step exceedance probability. Simulations of a linear two-CSTR-in-series process and of a nonlinear unicycle navigating past an obstacle under Student-$t$ disturbances validate both approaches and demonstrate reduced frequencies of safety constraint violations.

Independent control of tonescale regions (e.g., shadows, highlights) is essential for painters, photographers and cinematographers to bring 2D images to life. In image manipulation software this is most directly addressed by color grading modules, which use intensity thresholds to segment distinct illumination regions for local manipulation. In this work we develop an open source color grading tool and use it to annotate a large dataset of video frames with tonescale region thresholds. Using these thresholds we conduct modeling experiments with strategies based on both practitioners' conventional wisdom and machine learning. Results show that K-nearest neighbors is an effective prediction strategy, outperforming state-of-the-art end-to-end methods for image enhancement. This outcome demonstrates the benefit of focusing on a compact set of core parameters when modeling creative stylization processes. Our adaptive color grading interface and data are available at https://github.com/SamsungLabs/adaptive-color-grading.

Field robotics missions often require physical samples to be returned to laboratories for analysis, making path planning inherently load-aware and order-dependent as accumulated samples increase payload and traversal energy costs. In single-robot Load-Aware Informative Path Planning (LIPP), this rigidly couples sensing with hauling: a solitary robot must transport every collected sample, forcing frequent depot returns that severely restrict its spatial coverage. Heterogeneous multi-robot teams can overcome this bottleneck by dividing labor—enabling high-precision samplers to collect while high-capacity carriers handle transport. However, this introduces a complex coordination challenge regarding when, where, what, and to whom handoffs should occur on top of the LIPP problem. To address this tightly coupled problem, we introduce Multi-Agent LIPP (MA-LIPP), which enables teams to cooperate through asynchronous "dead drops," allowing one robot to deposit samples for another to retrieve later without requiring synchronous rendezvous. We formulate MA-LIPP as an exact Mixed-Integer Quadratic Program (MIQP) alongside a scalable Pairwise Large-Neighborhood Search (LNS) heuristic for complex real-world applications. The heuristic matches exact optima in $95.5\%$ of certified cases and reduces weighted posterior variance by $16.1$–$19.8\%$ relative to a sequential baseline on larger instances of up to 12 robots, providing a robust framework for cooperative physical-sampling missions.

Off-target protein binding is a major source of adverse effects for small-molecule drugs, yet most structure-based molecular design methods focus on generating selective compounds de novo rather than improving the selectivity of existing, well- characterized drugs. We introduce specificity optimization (SpecOpt), a molecular design task that seeks constrained structural modifications to an existing compound that increase its binding preference for an intended target over known off-targets while preserving its structural identity and drug-like properties. To enable systematic evaluation, we construct a ChEMBL-derived benchmark from compound-target interaction data, identifying intended targets through curated drug-mechanism annotations and off- targets through measured activities. We then develop an agentic framework that docks each compound against its intended target and off-targets, compares the resulting poses through residue-aware atom-protein contacts, and provides these differential interactions to a large language model to propose targeted structural modifications. Candidates are retained only if they satisfy molecular similarity, ADMET, and target-off-target docking selectivity criteria. On 915 compounds, the agent improves the target- off-target binding gap for 84.8% of compounds, shifting the mean gap from -0.72 to +0.47 kcal/mol while maintaining a mean Tanimoto similarity of 0.72 to the starting compounds. Ablation studies identify residue-specific contact information as the critical optimization signal: replacing residue identities with binary contact indicators eliminates improvement on all 29 ablation compounds. These results establish SpecOpt as a distinct molecular design problem and demonstrate residue-aware differential interactions as an effective signal for improving the specificity of existing compounds.

Integrating diverse data modalities — such as clinical notes, laboratory results, and medical imaging — is essential for advancing clinical decision-making. While Large Language Models (LLMs) have shown remarkable performance in processing unstructured clinical text, their limited capacity to incorporate non-text modalities hinders their broader utility in healthcare applications. Here, we introduce M2G-LLM (Multimodal MedGraph-LLM), a novel framework that enhances LLMs with multimodal integration and alignment via Graph Neural Networks (GNNs). Our approach models temporal relationships between patient visits, propagates information across clinically similar patients, and aligns heterogeneous data sources to construct enriched multimodal context vectors. These vectors are injected into the intermediate layers of the LLM, enabling joint reasoning over textual and non-textual modalities. We evaluate M2G-LLM on the MIMIC-IV and MIMIC-CXR datasets, demonstrating improvements in clinical prediction tasks over strong baseline models. Our results highlight the promise of combining the language understanding of LLMs with the relational reasoning capabilities of GNNs for comprehensive, multimodal healthcare analysis.

Artificial intelligence (AI) is increasingly embedded within scientific computing workflows that combine simulation, data processing, optimisation, visualisation and experimental or observational components. Learned models may serve as explicit workflow components, retain persistent state and, in adaptive settings, influence subsequent computation. Existing work has characterised scientific workflow management systems, dynamic and steered workflows, AI–HPC coupling motifs and the machine-learning lifecycle, although these areas are often discussed separately. This review brings them together from a systems perspective. We distinguish conventional scientific workflows, machine-learning pipelines, AI-coupled high-performance computing (HPC) workflows and broader automated research workflows, and propose a continuum describing the depth of AI participation from a computational stage to co-adaptive workflow control. The associated systems requirements are organised around five concerns: control and orchestration; compute and execution; data and model state; reproducibility and provenance; and governance and assurance. Representative systems and applications include AI-steered molecular simulation, drug and materials discovery, simulation–surrogate coupling and distributed self-driving laboratories. Workflow-level evaluation is considered in terms of scientific progress, execution cost, data movement, resource use, resilience and decision traceability. We conclude by identifying open problems in dynamic workflow representation, state-aware recovery, heterogeneous scheduling, interoperable data planes, model-mediated decision provenance and reproducible adaptive execution.

Most artificial neural systems are built to map given inputs to outputs. Adaptive agents face a prior problem: they must act without enough evidence, seek encounters with the world, and revise behavior when evidence appears. We propose another starting point for intelligent neural networks: proactive search without signals, curiosity at its most basic. We ask whether it can come from a minimal untrained circuit. The spiking unit studied here inverts its response to input: with no signal in its window it fires faster; once signals arrive it switches to a slower, inverted regime. Search needs three or more such oscillators in counter-phase, each reading the same input in a different time window. With no training, supervision, parameter tuning, or controller, the composite switches on its own between exploratory spiral search and exploitative tracking, finding both first-degree symmetry and second-degree groups. The switch comes from temporal disagreement between its fast and slow readings of the same signal. We view the circuit as evolutionarily trained: its abilities come from structure, not experience. Ablation over 63 configurations and 63,000 trials shows the switch needs both temporal staggering and counter-phase opposition, neither enough alone: the behavior is emergent, not programmed. The spiral persists at zero rotational diffusion, so it is structural, and degrades gently under perturbation. More oscillators improve spiral regularity but cut resource capture, so the smallest sufficient circuit wins. We propose that this principle underlies search in simple organisms, navigation and decisions in complex ones, and, being so simple and common, goes unnoticed unless you strip the logic bare. Eventually, networks of such proactive primitives may offer another foundation for AI architectures that explore our world rather than merely predict the next symbol in a sequence.

Voxel-based volumetric mapping is fundamental to 3D reconstruction, yet fixed-resolution grids remain inherently inefficient - wasting memory in uniform regions and losing detail in complex ones. Existing adaptive methods, such as MAP-ADAPT, partially address this by varying resolution based on geometry and user-defined semantic class lists, but these heuristics require expert tuning, lack generalization to unseen objects, and provide no explicit mechanism to control memory usage. We propose an adaptive framework that refines voxels based on semantic entropy, which captures label uncertainty, together with geometric curvature and texture richness as scene complexity cues, yielding principled resolution allocation without reliance on semantic taxonomies. To make the accuracy-memory trade-off explicit and user-controlled, we further introduce a reinforcement learning agent that learns voxel subdivision policies under a user-specified target memory budget, replacing hand-tuned thresholds with a single intuitive control parameter. The resulting multi-resolution TSDF achieves higher geometric accuracy, better semantic consistency, and improved memory-accuracy trade-offs compared to MAP-ADAPT and fixed-resolution baselines on both synthetic and real-world datasets. Our code and models are available at https://github.com/alpayozkan/UnRL.

Entanglement cost is the asymptotic rate of Bell pairs required to prepare a quantum state. Its regularized definition has made exact evaluation difficult, even for isotropic states (i.e., depolarized maximally entangled states). In this work, we determine the entanglement cost of every qubit isotropic state and, more generally, every two-qubit Bell-diagonal state. Our proof uses a family of suitably tuned qubit semigroups to transform a general log-Sobolev entropy bound into supporting lines of Wootters' function. A recent exact tensorization theorem of Dong et al. [arXiv:2606.17729] extends these bounds to arbitrarily many copies, yielding a general lower bound on entanglement cost. Matching this bound with the entanglement of formation also determines the exact cost for a broader class of two-qubit states. For qudit isotropic states, we derive another general lower bound from the full depolarizing $2\to 3$ norm. This bound substantially improves on the PPT-relative entropy of entanglement and nearly matches the entanglement of formation, leaving a gap of at most $2\%$ of $\log d$ that vanishes as $d$ grows. Finally, the exact qubit results and qudit bounds extend to the entanglement cost of preparing quantum depolarizing channels under both parallel and sequential strategies.

White men without an undergraduate degree receive the lowest average ratings among eight gender-race-education groups in controlled large-language-model evaluations of credit, hiring, and rental applications. We conduct full-factorial vignette experiments with 18 models from 12 developer groups, varying gender, race, age, citizenship, and education while holding stated financial or occupational circumstances constant within each setting. Each model evaluates all 32 profiles ten times per setting, yielding 17,280 ratings. Averaging over models, age, and citizenship, ratings for White men without degrees are the lowest among the eight groups, at 75.87 in credit, 92.71 in hiring, and 86.62 in rental housing on a 0-100 scale. Black women with degrees receive the highest average ratings, with corresponding gaps of 2.94, 3.66, and 3.56 points. Separate attribute effects favor women, Black applicants, and degree holders in all three settings. White men without degrees have the lowest or second-lowest mean in 46 of 54 model-scenario combinations (85.2%). This pattern connects to evidence of growing economic and health vulnerabilities among White men without degrees, highlighting a group whose disadvantages can be obscured by broad racial or gender categories.

Valid Model Context Protocol (MCP) messages do not guarantee that an agent framework preserves the distinctions downstream software needs. We present a differential testing method that follows a fixed tool result through each framework's public interfaces and checks explicit consumer requirements. Applied to 18 designed fixtures in four pinned Python integrations, the method identifies 13 unique fixture-task divergences involving structured values, declared errors, and rich content. Google ADK satisfies every primary contract in its observed path; the other integrations show interface-specific changes or an execution failure. Whether a change matters depends on the consumer: treating absent optional fields as equivalent to null explains most of OpenAI's strict rich-content failures. In an exploratory replay, parsing JSON text recovers more structured values but also returns incorrect values and values for fields absent at the source. Documented settings help in conflicting-output stress cases without restoring a separate structured slot. Fault challenges expose an oracle weakness and test its repair on fresh cases. The study provides reproducible, interface-specific evidence from controlled cases, not production failure rates or model-behavior measurements. Its practical implication is that post-wire regression tests need to specify both the information required and how the consumer reads it.

We study statistical estimation on implicit weighted similarity graphs presented as node-arrival streams. Previous work~\cite{LZ26b} obtained constant-pass, sublinear-space algorithms for several basic statistics of such graphs, including the diversity index $\mathsf{DI}=\sum_i d_i^{-1}$ and the degree moments $M_p=\sum_i d_i^p$ for $p>0$, together with their associated sampling problems $L_{\mathsf{DI}}$ and $L_{M_p}$. We settle three questions left unresolved by the previous work. First, we show that any two-pass streaming algorithm that $1.1$-approximates $\mathsf{DI}$ requires $\Omega(n)$ bits of space; combined with the three-pass $\tilde{O}(\sqrt n)$-space algorithm of~\cite{LZ26b}, this establishes a sharp transition between two and three passes for $\mathsf{DI}$. Second and third, for every fixed $c>0$ we give a three-pass $n^{-c}$-perfect $L_{\mathsf{DI}}$-sampler using $O_c(\sqrt n\log n)$ words and a two-pass $n^{-c}$-perfect $L_{M_p}$-sampler using $O_{p,c}(n^{1-1/(p+1)}\log n)$ words. Matching lower bounds up to polylogarithmic factors show that both the number of passes and the polynomial dependence on $n$ are optimal.

Large language model (LLM) judges are often used as rewards to train policies on objectives that deterministic verifiers cannot capture. However, these rewards are often collapsed to pass/fail ({0, 1}), which reports the verdict but not how well a response met each criterion. We model each pass/fail verdict as a score on an unreported scale, compared with one cutoff. A stretch of that scale moves every score proportionally toward or away from the cutoff, but never across it, so no verdict changes. A policy is therefore free to apply any stretch without changing anything the panel reports. Under a joint-Gaussian model, a third grade adds a second threshold and removes this affine stretch ambiguity. On MATH and SciBench outputs from one seven-criterion judge, all 14 constructed criterionwise stretches were invisible after binarization but visible with three grades. At $n=1{,}024$, a test given both population laws had at least 96.5% power at a $1.5\times$ stress. Retaining grades closes one blind spot created by binarization, but verdicts alone remain insufficient as some changes are still indistinguishable from genuine improvement. These include arbitrary within-grade changes and fixed-covariance, loading-aligned mean shifts – the signature of a sycophancy-shaped lift the panel reads as competence. The shared-factor reference approximation fit MATH and SciBench but not HealthBench, delineating its empirical scope. We recommend keeping at least three grades (for example, asking the judge whether each criterion is fully, partially, or not met and rewarding {0, 0.5, 1}), and externally validating gains along the remaining direction, which no finer scale removes.