Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Saturday 29 August
Video traffic constitutes a significant share of global web traffic. To reduce its volume, video codecs have been developed and continuously improved. While the industry has achieved substantial progress in traditional video coding, neural video codecs (NVCs) have recently emerged as a new approach that applies deep learning to video compression. This creates new challenges for compression quality assessment, which is essential for the further development and improvement of such codecs. In particular, it is important to evaluate the novel temporal compression paradigms introduced by NVCs. In this work, we present a large-scale subjective dataset of videos compressed with both neural and traditional video codecs. The subjective scores were collected through crowd-sourced pairwise comparisons. The proposed dataset provides a valuable resource for the development and benchmarking of video quality metrics tailored to neural video codecs. The dataset is available at the following link: https://videoprocessing.github.io/nvc-dataset-benchmark
Strongly Correlated Electrons
Collinear and noncollinear antiferromagnetic ordering in a highly frustrated decorated square kagomé lattice antiferromagnets of the nabokoite family
Nabokoite family compounds ACu$_7$(TeO$_4$)(SO4)$_5$X (A=Na, K, Rb, Cs; X=Cl, Br) host frustrated 2D square kagom'{e} lattice layers decorated by additional inter-layer magnetic ions. We study magnetic order in nabokoites with multi-frequency electron spin resonance spectroscopy and thermodynamic measurement (specific heat, magnetization and dielectric permittivity). Our study reveals that the choice of the low-temperature ground state is qualitatively different in light-alkali-ion (K, Na) and heavy-alkali-ion (Rb, Cs) compounds. Heavy-alkali-ion nabokoites order in conventional collinear antiferromagnetic pattern with easy-axis anisotropy. The parameters of the ordered antiferromagnetic state are very close for all heavy-alkali-ion subfamily. Light-alkali-ion members of nabokoite family demonstrate much more complicated route to the ordered state: firstly, a ferroelectric transition at 25-90K lifts the frustration and thus pre-cooks the low-temperature ordering; secondly, an unusual noncollinear magnetic order develops via two-step phase transition with first transition temperature $T_{c1}\simeq 5-6$K and the second transition at $T_{c2}\simeq 3-4$K. Noncollinear order is evidenced by observation of characteristic non-Larmor antiferromagnetic resonance mode. Spin dynamics of light-alkali-ion nabokoites is characterized by three zero-field magnon gaps and two spin-reorientation fields, the values of magnon gaps and critical fields are quite different for different compounds. The finite-size cluster modeling of pyramidal structural block of nabokoite structure combined suggests that the critical closeness of the nabokoite exchange coupling parameters to the border-line between the different quantum ground state of pyramidal building block of nabokoite structure could be the clue to the choice of qualitatively different ordered state in light- and heavy-alkali-ion nabokoites.
Read the full abstract
High Energy Physics (Theory)
Inequalities and Positivity in Modular Flowed Entanglement Entropy
The modular Hamiltonian plays an important role in quantum information theory, quantum field theory, and the AdS/CFT correspondence. In this paper, we study a specific dynamical setup: starting from the tensor product of the reduced density matrices of a reference state, we evolve this product state under the modular flow generated by the reference state itself. We then investigate the resulting change in the entanglement entropy of a given region, which we call the modular entanglement shift (MES). Unlike evolution under a fixed quantum channel, this modular evolution obeys no general principle requiring the MES to have a definite sign. For a bipartition into subsystems A and B, we establish an exact entropy balance: the sum of the MESs for A and B equals the mutual information generated along the orbit. Consequently, their sum is nonnegative, even though either individual shift may be negative. For two disjoint intervals in the vacuum of a two-dimensional conformal field theory, however, we prove that a global conformal involution exchanging the intervals forces the two MESs to be equal. It follows that each MES is nonnegative along the modular flow. We establish this result both using a conformally transported regulator and intrinsically in terms of Araki relative entropy. Finally, we identify the modular-flow orbit with a Connes cocycle orbit and discuss possible applications to the AdS/CFT correspondence.
Read the full abstract
Human-Computer Interaction
Understanding Behavioral Dark Patterns of High BMI Individuals
Understanding how everyday behaviors influence body weight is essential for designing effective and personalized health interventions. Existing studies largely rely on self-reported questionnaires or limited sensing modalities, making it difficult to capture the temporal dynamics of daily behavior. In this work, we analyze the DiversityOne dataset, comprising four weeks of passive smartphone sensing and ecological momentary assessments collected from 453 university students across eight countries. We extract behavioral features spanning dietary habits, physical activity, screen time, and smartphone usage, and investigate their associations with self-reported Body Mass Index (BMI). Beyond feature-level analysis, we employ Hidden Markov Models (HMMs) to uncover latent behavioral patterns. Our analysis reveals that higher BMI is associated with more frequent consumption of soda, alcohol, and processed meat. We further reveal that overweight and obese individuals spend longer periods in food delivery apps and are more likely to transition back to unhealthy eating and drinking routines after starting to exercise. In contrast, normal-weight individuals lead a more balanced lifestyle. These findings highlight key behavioral patterns that make weight loss particularly challenging.
Read the full abstract
Computation and Language
When to Adapt: Conditional Memory Adapters for Retention-Preserving Domain Specialization
Large language models deployed in specialized domains must improve in-domain performance without sacrificing general capabilities. Existing parameter-efficient fine-tuning methods are typically always on: their learned perturbations are applied to every input, which can degrade out-of-domain (OOD) performance. We propose Engram Adapter, a framework that repurposes pretraining-time conditional memory as a post-hoc adapter for frozen LLMs. It uses multi-channel matching over local n-gram patterns with explicit occupancy tracking as a lightweight selectivity prior, making residual injection more likely on in-domain inputs while a learned scalar gate suppresses incoherent OOD retrievals. We evaluate on Qwen3-4B and Qwen3-8B with AG-News and MedMCQA as adaptation tasks and OOD benchmarks spanning reasoning, translation, code generation, and legal reasoning. Engram Adapter improves in-domain accuracy while preserving 99.4%–100.1% of average OOD performance; on LegalBench it slightly exceeds the frozen base model on average, whereas comparable always-on baselines degrade sharply. Mechanistic analyses show that although OOD activations are non-zero, gate and projection attenuation reduce residuals to approximately 0.08% of hidden-state norm, yielding small KL drift and negligible accuracy change. These results suggest conditional activation is a promising route toward modular, retention-preserving domain specialization over frozen backbones.
Read the full abstract
Computation and Language
StageWell: A Process-Aligned Chinese Corpus for Positive-Psychology Support Dialogue
Positive psychology dialogue aims to support emotional distress and positive resource building, requiring models to produce not only empathetic replies but also coherent progression through a multi-turn support process. Existing resources often reduce supervision to turn-level strategies or holistic preference labels, leaving process position, support function, and local repair targets implicit. We introduce StageWell, a process-aligned Chinese corpus for positive psychology dialogue, together with HQS, a structured protocol for data construction and evaluation. StageWell organizes support into a six-stage support process and uses a multi-agent whole-dialogue rewriting workflow to construct 12,445 SFT instances, 1,849 DPO preference pairs, and a GroundTruth subset of 120 expert-revised dialogues and 977 QA pairs. Guided by HQS, DPO pairs are built as process-localized repairs: flawed model outputs are used as rejected responses, and targeted rewrites under the same context and stage constraint are used as chosen responses. Across four 9B-14B open-source LLMs, this supervision yields robust gains in process control, response quality, and safety. Averaged across models, BERTScore improves by 0.037, Q-Overall increases by 1.32 points, S-exact increases by 0.236, and the H-critical rate decreases by 0.167. These results highlight the value of modeling supportive dialogue as a structured multi-turn support process rather than as single-turn response generation.
Read the full abstract
Mesoscale and Nanoscale Physics
Heavy-Hole–Light-Hole Mixing and Spin–Photon Coupling in Germanium Flopping-Mode Qubits
Intrinsic spin–orbit interaction in germanium provides electrically active flopping-mode qubits without requiring magnetic-field gradients to generate spin–charge hybridization. We study double quantum dots (DQDs) with a multiband Luttinger–Kohn framework that retains heavy-hole (HH) and light-hole (LH) states on equal footing. Modeling vertical confinement with a finite confinement potential along the growth direction brings selected subbands into a regime of appreciable HH–LH mixing. This admixture activates electric-dipole spin-coupling through LH–LH and LH–HH transitions, complementing the conventional HH–HH pseudospin-flip resonance. For the parameter sets considered, these additional channels can produce larger spin–photon figures of merit than the HH–HH transition. HH–LH mixing thus supplies a tunable ingredient for qubit–cavity coupling in germanium hole-spin qubits and extends the operating regimes available to circuit-QED architectures.
Read the full abstract
Computation and Language
Padārtha: Ontology-Grounded Fine-Grained NER Benchmark for Classical Sanskrit
Annotation schemas are not neutral. When applied to classical literature, tag sets developed for modern journalistic texts impose source-culture definitions on texts they were never designed to describe. We instead ground a schema in the tradition of the text itself introducing Padārtha, the first ontology-grounded fine-grained Named Entity Recognition (NER) benchmark for Sanskrit, built on the Mahābhārata epic. Our tag set derives from Nyāya-Vaiśesika, a classical Indian ontological system, yielding 18 fine-grained categories organized under 10 ontological nodes and mapped onto five standard coarse tags, ensuring interoperability with existing benchmarks. Expert annotators label over 12.6K entries from a scholarly index of named entities, linked to corresponding mentions in the Mahānāma corpus, producing fine-grained annotations for 108,335 entity mentions across 73,632 verses, along with a 5,000-verse expert-verified test set sampled to stress rare mentions. We present the first systematic benchmarking of generative NER against traditional architectures for Sanskrit, finding that fine-tuned generative models perform comparably to task-specific systems. However, all systems show a sharp decline from coarse to fine granularity and struggle with out-of-entity mentions unseen during training. The limitation is not due to data scarcity alone, as fine-tuned models recall unseen entities far worse than seen ones and tend to default to the majority sense under lexical ambiguity.
Read the full abstract
Strongly Correlated Electrons
Experimental observation of decoupled spin subsystems in decorated square kagomé lattice magnets of the nabokoite family
The square kagomé lattice (SKL) offers a model platform for investigating geometric frustration in 2D systems. Nabokoite-family compounds \nabok{A}{X} (A=Na, K, Cs, Rb and X=Cl, Br) extend this physics to a 3D network, where 2D SKL layers are decorated by interlayer spins. Using electron paramagnetic resonance (EPR), we demonstrate a dramatic splitting of this complex exchange network into two virtually decoupled spin subsystems: absolute calibration of the electron paramagnetic resonance (EPR) absorption reveals that only a fraction of all copper spins in nabokoites is EPR-active and this fraction of the spins orders at the Néel point. Comparison of the EPR absorption and static susceptibility indicates that contribution of the EPR-silent spin subsystem to total magnetic susceptibility decreases on cooling. This direct observation of coexisting magnetic order and possible spin-liquid dynamics within a single compound challenges conventional models of unified exchange networks in decorated frustrated lattices.
Read the full abstract
Computer Vision
Test-Time Scaling for Video Diffusion Models via Diagnosis-Guided Candidate Recycling
Recent video diffusion models have achieved remarkable generation quality, but high-fidelity results still largely depend on closed-source systems or costly large-scale infrastructure. Test-time scaling (TTS) offers a training-free way to improve lightweight generators by spending additional inference compute, yet existing methods mostly remain within a noise-search paradigm: they sample, select, or perturb denoising trajectories and discard low-scoring candidates after expensive generation. This generate-and-discard process wastes not only computation but also the partial motion, layout, or appearance structure already encoded in recoverable samples. We present GEARS (Guided Editing for Adaptive Recycling Search), a training-free framework that introduces {diagnosis-guided candidate recycling} into video TTS by turning such candidates into editable priors through a generation-evaluation-editing loop. GEARS consists of two collaborative components. The Stage-Aware Scheduler determines what to repair, when to repair it, and which candidates should be preserved, recycled, or discarded. The Candidate Recycler diagnoses recoverable failures from keyframes and multi-dimensional reward feedback, derives candidate-specific repair prompts, and repairs the corresponding candidates through manifold-aware latent SDEdit. The repaired candidates are recycled into the search pool, creating refinement paths beyond standard noise perturbation while preserving useful structure. Under matched NFE budgets, GEARS consistently outperforms existing video TTS methods on VBench, bringing a 1.3B model to a total score comparable to a 14B counterpart, and ablations verify the necessity of adaptive scheduling, diagnosis-conditioned editing, and manifold-aware re-denoising. Code is available on GitHub.
Read the full abstract
Mathematical Physics
Simulating passive scalar advection in rough Kraichnan flows
We present a systematic Eulerian study of passive scalar advection in rough two-dimensional Kraichnan flows, covering the full range of velocity roughness exponent $h\in(0,1)$. The advection–diffusion equation is integrated directly using a pseudo-spectral method at resolutions up to $2048^2$ grid points, with a frozen-noise Runge–Kutta scheme consistent with the white-in-time construction of the carrier flow. The simulations recover the duality between advected scalar and advecting flow—the smoother the carrier, the rougher the scalar—together with the predicted scaling laws for the second-order statistics. Once the scaling range is properly identified, the fourth-order flatness anomaly is measured across the whole range of $h$, in agreement with the Lagrangian estimates of Frisch et al. (1999) and with the perturbative predictions of Bernard et al. (1998) and Pumir et al. (1997). Benefiting from the fact that our Eulerian simulations give direct access to the full scalar field, we also examine the probability density functions of scalar increments and the corresponding higher-order statistics, which show systematic departures from Gaussianity and from log-normality, with the strongest deviations manifesting for intermediate values of $h$. A central outcome of this work is a systematic account of how the simulation parameters, in particular the molecular diffusivity, must be adjusted with $h$, providing practical guidelines for reliable simulations; we further show that the residual deviations from the theoretical scaling laws are quantitatively accounted for by the finite spectral representation of the carrier flow. Our analysis also serve as a numerical baseline for simulating more realistic extensions of the Kraichnan model, where the carrier flow is coupled with a Gaussian multiplicative chaos and theoretical developments are limited.
Read the full abstract
Algebraic Geometry
Ample vector bundles on non-proper schemes
We solve two open problems on ample vector bundles posed by Hartshorne in 1966. We prove that tensor products of ample vector bundles on schemes of finite type over an algebraically closed field are ample in arbitrary characteristic, extending Hartshorne's characteristic-zero result and Barton's projective positive-characteristic theorem. More generally, let f: X -> S be a morphism of schemes. Then we prove that tensor products of f-ample vector bundles are f-ample. Moreover, if E is an f-ample vector bundle of rank r>0 and W is a finite locally free polynomial GL(r,S)-module of positive rank with W_0 = 0, then E(W) is f-ample. In particular, Gamma^n E is f-ample for every n>0. Finally, adapting a construction of Ejiri-Fujino-Iwai, we show that in every characteristic a smooth quasi-projective surface carries a non-ample extension of an ample line bundle by itself.
Read the full abstract
Formal Languages and Automata
The Emptiness Problem for Quantum Finite Automata with Classical States
Quantum Finite Automata with Classical states (QFACs) are nondeterministic finite automata over a finite alphabet of quantum operations. We study expressiveness of this model on finite words and the corresponding emptiness problem. We show that regular languages are incomparable with those definable by Quantum Finite Automata (QFAs) and that both are strictly subsumed by QFAC-definable languages. We show that the emptiness problem for a QFAC can be reduced to the emptiness of the language intersection of a QFA and a finite automaton. This intersection is known to be decidable for strict thresholds but undecidable for non-strict cases. Furthermore, we consider the problem for flat QFACs, a restriction where the underlying automata contain no nested loops, and relate it to the higher-dimensional orbit problem, a long-standing open challenge in dynamical systems. Finally, we propose a sound and semi-complete witness searching procedure to verify the non-emptiness of one-loop QFACs, which are sufficiently expressive to represent some prominent quantum algorithms, such as Grover's search and quantum random walks.
Read the full abstract
Differential Geometry
The Faraday Form of Conformal Products: Weyl Curvature and Four-Dimensional Einstein Rigidity
Let $(M^n,g)$ carry a conformal product structure whose adapted Weyl connection $D$ preserves orthogonal distributions of ranks $p,q\ge2$. We express the Faraday form $d\theta$ directly in terms of the Weyl curvature of $g$. If $S$ is the orthogonal involution determined by the splitting and \[ \mathcal K_W(X)=\sum_i W_{X,e_i}(Se_i) \] for any local orthonormal frame, then \[ (d\theta)^\sharp = \frac{n-2}{4(p-1)(q-1)}[\mathcal K_W,S]. \] No Ricci-curvature assumption is required. The same curvature calculation gives a companion formula for the symmetric part of $\nabla\theta$, \[ S\Theta^s-\Theta^sS+\theta^\sharp\wedge S\theta^\sharp = \frac{p-q}{n-2}(d\theta)^\sharp +\frac1{n-2}[S,\Ric^\sharp], \qquad \Theta=\nabla\theta. \] When the Ricci tensor is block diagonal with respect to the product splitting, these formulas lead to explicit identities for the two components of the Lee form. In dimension four, the rank-$(2,2)$ splitting determines an ambi-Hermitian pair. The classical ambi-Hermitian curvature identities imply local closedness when the Ricci tensor is invariant under both complex structures; we also recover this conclusion from the mixed Weyl trace above. Finally, if $(M^4,g)$ is compact and Einstein, no specialness assumption is needed: $D=\nabla^g$. Hence the universal cover is the product of two simply connected surfaces with the same constant Gaussian curvature.
Read the full abstract
Atomic Physics
Controlling light shifts in chip-scale atomic beam clocks
Chip-scale atomic beam clocks are being investigated to extend the range of clock stability achievable in low-power timing applications. Here, we demonstrate a centimeter-scale, Ramsey coherent population trapping (CPT) clock based on a microfabricated Cs atomic beam device and investigate the interplay between light shifts and Doppler shifts that determines its leading clock systematics. We show that these shifts exhibit competing dependencies on CPT light parameters, leading to “doubly-insensitive" operating points where the clock frequency is simultaneously insensitive to laser frequency and power. We further demonstrate a method for controlling key clock shifts using spectroscopic signatures from the atomic beam that is compatible with fully-integrated operation. The clock achieves a fractional frequency stability of $2 \times 10^{-10}$ at $1~\textrm{s}$ and sub-$\mu$s drift over nearly $17~\textrm{hours}$, with leading CPT light systematics controlled below the $10^{-12}$ level.
Read the full abstract
Differential Geometry
Powers of the Dirac Operator on $S^1\times S^{n-1}$
We give explicit formulas for the spectra of the intertwinors on the spinor bundle over $S^1 \times S^{n-1}$, $n$ even, with the standard Lorentzian metric $g=-g_{{}_{S^1}}+g_{{}_{S^{n-1}}}$. As a special case, we construct conformally covariant differential operators of all odd orders with the leading term a power of the Dirac operator.
Read the full abstract
Robotics
SGE: Semantically-Guided Exploration for Unstructured Environments via Image-Space Waypoint Sampling
This work introduces Semantically-Guided Exploration (SGE), a modular exploration framework for ground vehicles that integrates pixel-level semantic segmentation into sampling-based waypoint selection and receding-horizon route optimization. Unlike conventional geometric exploration methods, SGE evaluates candidate exploration goals directly in the image space using a semantic-aware utility function that accounts for terrain traversability, obstacle proximity, objects of interest, and depth-based exploration reward. Sampled waypoints are projected into 3D and ordered through a real-time Traveling Salesman Problem (TSP) formulation, enabling receding-horizon goal selection. To address real-world navigation uncertainty, the framework introduces mechanisms, including temporary taboo regions to handle navigation failures and a graph-based relocation strategy for efficient backtracking across explored areas. We evaluate SGE in standardized simulation benchmarks against state-of-the-art exploration planners and demonstrate competitive performance in volumetric coverage, while enabling semantic task biasing that cannot be achieved by purely geometric methods. The framework is further validated through real-world experiments using multiple robotic platforms in indoor campus buildings and in limestone and coal mines. Results show consistent performance and adaptability across platforms and domains.
Read the full abstract
Computation and Language
All You Need Is Non-Commutative Words
We represent lexical tokens as unitary matrices and encode each sentence as their ordered product. The noncommutativity of matrix product captures word order without positional encodings (PEs). The same algebra yields several capabilities, including antisymmetric self-attention with no query, key, or value projections, and parallel composition of variable-length text chunks at a reduced attention cost. Furthermore, it provides a canonical-coset readout layer that encodes all true unitary degrees of freedom compactly, while supporting continual learning through nested group extensions that enlarge the operator space with each new task preserving prior representations exactly. Across standard text-classification benchmarks, the method matches or exceeds bag-of-words baselines. Achieving higher accuracy on IMDB and comparable performance on AG News. Notably, this is accomplished by replacing the conventional $\sim$30,000-dimensional vocabulary space with a dense, 64-parameter real-valued encoding, highlighting the expressive efficiency of our parameterization.
Read the full abstract
Computer Vision
Hyper3-CLIP: Hierarchy-Conditioned Hyperbolic Vision-Language Training
CLIP-like vision-language models (VLMs) trained with contrastive objectives learn strong global image-text representations, but their Euclidean embeddings and global pooling fail to encode relational structure such as part-whole and parent-child relations. Hyperbolic VLMs address this gap with entailment-based objectives, and text-conditioned variants improve fine-grained alignment through sentence- and phrase-level queries. However, these two lines of work remain separate: hyperbolic VLMs use static image and region features, while query-conditioned methods lack hierarchical geometric structure. We present Hyper3-CLIP, a hierarchy-conditioned hyperbolic VLM that combines global, local, and global-local contrastive learning with query-conditioned visual pooling. To train the model, we construct lightweight query hierarchies from text, comprising full captions, sentence fragments, localized part descriptions, and extracted phrases. Each query conditions the pooling of visual patches, and the resulting representations support image-text, whole-part, and parent-child entailment losses. Query-conditioned pooling is active only during training. Hyper3-CLIP improves R@5 and R@10 retrieval on COCO and Flickr, as well as multi-label classification on VOC and COCO, while remaining competitive on hierarchy metrics. We also audit zero-shot prompt sensitivity under fixed prompt regimes and study the effect of the localized GRIT part budget used during training. Code is available at https://github.com/Hyper3Labs/hyper3-clip.
Read the full abstract
Representation Theory
Generalized Jantzen filtration, hyperbolic restriction, and characteristic cycles
In this paper we give a geometric description of the behavior of analytic intertwining operators between parabolically induced representations of $\mathrm{GL}_n(\mathrm{F})$, where $\mathrm{F}$ is a local non-archimedean field, in terms of hyperbolic localization functors of Braden. As a consequence, we can show that the image of an intertwining operator is always semi-simple and give a lower bound on the order of its pole, which is conjectured to be an equality as well as an upper bound in terms of the singular support of certain perverse sheaves. Moreover, we are able to reduce the conjecture of Lapid and Mínguez on the shape of irreducible subrepresentations of induced representations to a computation of characteristic cycles. The theory of mixed Hodge modules plays a crucial role in the proofs.
Read the full abstract
Artificial Intelligence
Formal Concept Analysis with Three Types of Negation
Classic Formal Concept Analysis (FCA) primarily focuses on the positive relationships between objects and attributes and does not have mechanisms for handling negation.To overcome this limitation, we introduce three types of negation concepts (contradictory negation, opposite negation, intermediary negation) into FCA.Based on the set SCOI and logic LCOI+PLCOI with these three types negation, we define formal context, Galois connection operators, formal concept and concept lattice with three types of negation,this leads to the proposal of a FCACOI: Formal Concept Analysis with contradictory negation, opposite negation and intermediary negation.For the reasoning in FCACOI, this paper focuses on attribute implication reasoning. Based on the logic LCOI+PLCOI and its semantics, we introduce the notion of ICOI-entailment as the semantic implication for attribute implication reasoning in FCACOI. Through ICOI-entailment, a connection is established between attribute implication reasoning in FCACOI and inference in the logic LCOI+PLCOI, it indicate that formally proven inference rules (theorems) in LCOI+PLCOI are valid in the attribute implication reasoning of FCACOI, LCOI+PLCOI provides a logical foundation for attribute implication reasoning in FCACOI. To illustrate the capability of attribute implication reasoning in FCACOI, we discuss its application in a concrete example. Moreover, we explore attribute reduction of the formal context in FCACOI, propose two research frameworks for attribute reduction from different perspectives, and compare their characteristics.We believe that, based on richer logic and semantics, FCACOI elevates FCA from a theory that describes affirmations to one that can describe affirmations and its contradiction(either this or that), opposition(extreme negation) and intermediary (transitional states between oppositions).
Read the full abstract
Software Engineering
Super Library Agent: Joint Generation and Maintenance of Multiple Applications Beyond the Single Codebase
Organizations often develop and maintain portfolios of related applications: independently deployable codebases that share substantial domain logic, interface patterns, or operational conventions. As LLM coding agents are increasingly used to generate and maintain such software, a naive application-by-application workflow duplicates shared logic across codebases and allows prolonged agentic maintenance to accumulate verbosity, dead code, and structural erosion. We introduce the Super Library Agent problem, where an agent sequentially generates a portfolio of N related applications while maintaining a shared Super Library of reusable cross-application components. A minimal sequential scaffold can in principle extract shared code and migrate applications to the evolving library, but in practice suffers from low extraction recall and fragile dependency migration. We address these failures with candidate-guided extraction over code chunk summaries, pre-extraction codebase consolidation, and context-aware migration using extraction traces and call-graph information. Across WebGen-Bench and PaperBench, our method preserves application functionality while significantly reducing redundancy and token footprint (verbosity, token length) over zero-shot, and avoiding the structural erosion introduced by naive library construction, with additional reductions in LOC and MDL. Our code is available at https://github.com/sbigstar0310/super-library-agent.
Read the full abstract
Materials Science
Evaluating LLM-based AI agents integrated with materials synthesis tools: the case of atomic layer deposition
This work provides an overview of the different strategies that can be used to evaluate the performance of AI models and agents based on large language models (LLMs) for materials synthesis. After providing a brief overview of the key technologies behind the current generation of AI agents based on LLMs, we summarize the different approaches to evaluating these models in the context of materials science and in particular on materials synthesis, with a specific emphasis on scenarios in which the models are directly integrated with experimental tools. We discuss evaluation strategies spanning knowledge and reasoning benchmarks, tool-use benchmarks, and closed loop benchmarks involving the interaction with experimental systems or realistic virtual tools. We use atomic layer deposition (ALD) as a case study, emphasizing how existing approaches in the literature both build from general approaches used beyond materials science and can be generalized to other materials synthesis techniques. Finally, we provide a practical evaluation framework to evaluate LLMs in the context of materials synthesis
Read the full abstract
Game Theory
Improving Randomized Metric Distortion to 2.1441
In metric social choice, voters rank candidates by their distances in an unknown metric space. A voting rule uses these rankings to select a candidate or a lottery over candidates, aiming to minimize the average distance to voters. Distortion measures the worst-case approximation ratio. While the best distortion of deterministic rules is $3$, prior work pins down the best distortion of randomized rules to $[2.1126,2.5]$. We improve the upper bound to $2.1441$, closing over $90\%$ of this gap. The proof introduces random-size stable lotteries, proves their existence, and derives the new bound through a potential argument. All the proofs were obtained using GPT-5.6-Sol with significant guidance from the author, who verified them and simplified the exposition.
Read the full abstract
Computation and Language
Detecting and Repairing Hallucinations in Retrieval-Augmented Generation
Language models increasingly answer questions by consulting retrieved documents rather than memory alone, a design now common in search assistants and enterprise knowledge tools. Grounding a model in retrieved text reduces unsupported statements but does not eliminate them, and a reader cannot tell a grounded sentence from an invented one. Most research on this problem stops at detection, yet flagging a faulty answer changes nothing for the person reading it, and little is known about which action should follow. Using RAGTruth, a benchmark whose unsupported passages are annotated by hand, we split each flagged answer into individual factual claims, check each against the retrieved source, and compare leaving the answer untouched with three repair strategies of increasing richness: deleting an unsupported claim, replacing it with source text, and rewriting it. Three language models from different families judge the 916 repaired answers. Every strategy reduces the proportion of answers judged to contain unsupported content, and all three judges agree on the ordering. Deletion achieves the largest reduction while retaining least of the original answer, at 64.3% of the text, whereas rewriting retains 80.1% and reduces least. Repair is not confined to faulty answers: 83.5% of answers annotated clean are edited too. The strategies occupy different points on a grounding preservation trade-off rather than forming a quality ranking, and choosing between them needs evidence about answer usefulness that automatic metrics cannot supply.
Read the full abstract
Computers and Society
The relationship between professional and general ethics in generative AI
Recent years have seen a growing discrepancy in the field of AI alignment: research and policy recommendations on AI ethics tend to assume a general set of ethical values, yet proliferating practice-specific uses of AI systems on the ground - in the legal, medical and translation domains, among others - have been effectively manifesting ethics of professional practice. This article begins by outlining the reasons why general and professional ethics are increasingly conflicted in contemporary AI systems, and by surveying how the research literature attests to, but has not yet resolved, this conceptual and practical challenge. We then conceptualize the main dimensions of AI models' decision-making in areas of professional practice, emphasizing professional ethics' hierarchically structured relationship with general ethics, and elaborating on the mechanisms through which they reach an equilibrium in situational contexts that involve conflict. It is through this equilibrium, we suggest, that certain professional ethics are prioritized over others and implemented in practice. We then show how our framework can be the basis for a systematic empirical assessment of AI models' professional ethics in various domains, identifying the nuances of the models' favored ethic by examining their production in a series of similar but not identical scenarios. Finally, we propose a formulation for how to intervene in and change AI models' favored ethics in professional practices - while noting the inherent dimension of subjectivity involved in both the evaluation and implementation of professional ethics in AI models.
Read the full abstract
Computation and Language
Learning Simple Test-Time Environments for LLM Web Agents
Large language model (LLM) agents have demonstrated remarkable proficiency in manually constructed environments, yet their performance frequently collapses when transitioned to complex real-world settings. Existing research largely attribute this degradation to the compositional generalization gaps in LLMs on combinations of multiple simple, well-structured environments. In this work, we propose that LLM web agents can learn simple environment observations at test time. Specifically, we introduce trial steps for agents to decompose a complex environment observation into sub-modules, and implement a label-free learning method, Test-Time Environment Decomposition (TTED), to adapt agent behaviors with experience during inference. Our empirical evaluations demonstrate the framework's efficacy across both synthetic and realistic benchmarks, showing (1) experience gains acquired within simpler sub-environments can be effectively composed to improve performance in the full one, and (2) test-time training on sub-environments can significantly enhance the compositional generalization of agents in real-world web automation tasks. We also provide key insights in the design of the label-free learning algorithm. As more complex environments are accessed by LLM agents, we believe learning environment decomposition skills at test time will be critical for robust real-world deployment.
Read the full abstract
Machine Learning
MEL: Coordinate-Preserving EEG Tokenization for fMRI Translation
Background and Objective: Predicting regional blood-oxygen-level-dependent (BOLD) activity from EEG provides a computational approach to studying electrophysiological-hemodynamic associations. However, changes in input representation, readout architecture, and ensembling can obscure the source of a prediction gain. We investigate the contribution of target-relative spectral estimation under matched readout conditions. Methods: MEL converts each preceding 16-second EEG history into calibrated lag-channel-frequency measurements. Four equal-dimensional representations are compared using matched Ridge and multilayer perceptron (MLP) readouts. The retrospective analysis includes 29 resting-state scans from 22 participants, participant-paired uncertainty estimates, aggregation and correspondence controls, leave-one-participant-out regression, and refitted models on an external Oddball cohort. Results: Morlet-Ridge achieved a mean scan-wise correlation of 0.508 across six regional targets and the global signal, compared with 0.432 for canonical Fourier features. The participant-paired mean difference was 0.069 (95% bootstrap confidence interval, 0.037-0.101), with improvements in 19 of 22 participants. The Morlet MLP prediction ensemble reached 0.491. Shared coordinate permutations preserved prediction after refitting, whereas sample-wise permutations degraded it. Canonical Fourier regression reached 0.377 under participant exclusion; external unseen-participant correlation was 0.026. Conclusions: Spectral estimation materially affected regional BOLD prediction without requiring additional nonlinear readout capacity. MEL provides an explicit measurement interface for controlled comparisons, but weak external results constrain its generalizability. Confirmation requires untouched participant-separated data and split-aware preprocessing.
Read the full abstract
Systems and Control
Learning neural controllers for nonlinear systems from data
This article addresses the problem of designing neural feedback controllers for unknown nonlinear systems. We propose an indirect data-driven framework that uses offline data to identify the system dynamics, upon which a neural feedback controller and a neural Lyapunov function are jointly synthesized. Input constraints are enforced by integrating a hard-saturation structure into the controller architecture. Robust synthesis conditions are derived to account for data perturbations during identification. Formal stability is certified by combining SMT verification with local Lyapunov analysis near the equilibrium. Numerical examples validate the effectiveness of the proposed framework.
Read the full abstract
Machine Learning
A Spectral Identifiability Threshold for Dissipative Rate Recovery from Truncated Liouvillian Spectra
Open quantum systems lose energy and phase coherence through different dissipative processes, but these processes can produce overlapping dynamical signatures. The Liouvillian spectrum summarizes how such a system relaxes, yet it is not obvious how much of that spectrum is needed to distinguish the underlying dissipation rates. We study this question for amplitude damping and dephasing in a six-qubit Lindblad model whose spectrum can be derived analytically. We retain only the slowest non-steady spectral modes and ask how many are required before each dissipative rate becomes recoverable. We show that population modes contain no dephasing information, which creates a lower bound of D = 2^n retained modes for uniform dephasing identifiability in the relevant rate regime. The measured recovery threshold reaches this bound at n = 4,5,6, while n = 3 remains above it. At n = 6, least squares achieves a mean joint absolute error of order 10^-9, compared with 4.355 x 10^-4 for four tabular learning methods. Robustness tests show that this advantage weakens when the spectra are perturbed and when a transverse field breaks the commuting structure. These results show that the amount and structure of retained spectral information can determine whether dissipative parameters are recoverable, independently of the estimator used. The present conclusions apply to noise-free simulator spectra rather than measurement-derived spectra.