Human-Computer Interaction
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 31 August
Extended Reality (XR) systems are often most useful when users are engaged in ongoing physical tasks, yet current interaction techniques still largely assume the hands are available. We present opportunistic body input, an interaction paradigm that redirects continuous XR control to whichever available body region remains free in the moment. To investigate how users naturally coordinate these spare-body movements, we conducted an elicitation study across six hand-busy scenarios. We found that while users' preferred spare body regions shift dynamically based on physical constraints, the resulting spontaneous movements share a consistent, low-dimensional kinematic structure organized around a dominant principal axis. Building on these findings, we present DOBI (Dynamic Opportunistic Body Input), a real-time XR technique that uses gaze to target a UI element, a brief trigger gesture to identify the recruited spare body region, and the region's subsequent motion to drive continuous 1D control. A 1D Fitts' law study establishes the baseline motor performance of this paradigm across four distinct body regions, achieving throughputs up to 2.62 bits/s with an overall 5.0% error rate, and a dual-task usability study shows that DOBI supports reliable, low-effort control (SUS = 84.2) while users remain engaged in realistic hand-busy activities.
Quantum Physics
Exact quantum splitting and the structure of finite algebras
For squarefree $f\in\F_q[x]$, Berlekamp's fixed-point algebra is $B\cong\F_q^r$, where $r$ is the number of irreducible factors. We give an exact quantum splitting procedure in a circuit model permitting efficiently computable, input-dependent single-qubit rotations. For a block with $s\ge2$ factors, a quadratic-character test (odd $q$) or absolute-trace test (characteristic $2$) is nonconstant with an exactly known probability $p_{q,s}\ge1/2$. Its value depends only on $q,s$, not on the hidden factors. An auxiliary-qubit reduction to success probability $1/4$ makes one Grover iteration exact. Thus factorization takes exactly $r-1$ splitting rounds and $O(n^3\log q)$ quantum $\F_q$-operations, with deterministic polynomial classical preprocessing and bookkeeping. The method also extracts primitive idempotents of arbitrary finite-dimensional commutative unital algebras given by structure constants in $O(n^4\log q)$ quantum $\F_q$-operations, staying over $\F_q$ throughout. Computing the radical and passing to the center extends it to the decomposition of $A/\Rad(A)$ into simple two-sided ideals, for an arbitrary finite-dimensional associative $\F_q$-algebra $A$.
Read the full abstract
Operator Algebras
Pointwise convergence of noncommutative ergodic averages along the primes
Let $\mathcal N$ be a von Neumann algebra equipped with a normal faithful semifinite trace, and let $\gamma$ be a trace-preserving automorphism of $\mathcal N$. We consider the ergodic averages along the prime numbers \[ A_N(x) := \frac1{|P_N|} \sum_{q\in P_N}\gamma^q(x), \qquad P_N:=\{q\leq N:q\ \text{is prime}\}. \] For every $1<p<\infty$, we prove a strong maximal inequality for $(A_N)_{N\geq2}$ on $L_p(\mathcal N)$ and that $A_N(x)$ converges bilaterally almost uniformly for every $x\in L_p(\mathcal N)$. The proof exploits the circle method and a noncommutative sampling principle. For the convergence result, Bourgain's commutative argument uses pointwise maximal functions and exceptional sets. These tools are not available in the noncommutative setting. Instead, we show that the tails of the ergodic averages tend to zero in $L_2(\mathcal N;\ell_\infty)$ and that the difference from the limit belongs to $L_2(\mathcal N;c_0)$. This gives the desired b.a.u. convergence, and provides a positive answer to one question left open in \cite{ChenHongWang+arXiv2024}.
Read the full abstract
Computational Physics
Cartesian tensor equivariant machine-learning force field for spin-dependent atomistic simulations
Magnetic materials exhibit an intricate coupling between atomic structure and spin degrees of freedom, posing a fundamental challenge for atomistic simulations across experimentally relevant length and time scales. Here we introduce HotPP-Spin, a spin-dependent extension of HotPP for magnetic machine learning interatomic potentials, built on Cartesian tensor equivariant message passing. Atomic magnetic moments are treated as explicit axial-vector degrees of freedom, while spatial-inversion and time-reversal parities are propagated through the tensor couplings. This construction provides a unified representation of exchange-dominated and spin-orbit-induced interactions without imposing predefined analytical interaction forms. A scalar spin-dependent potential energy surface yields energy-conserving atomic forces and magnetic effective fields through differentiation. Benchmarks spanning collinear magnetism, noncollinear magnetism, and spin-orbit-coupling-induced magnetic anisotropy show that HotPP-Spin accurately describes magnetic energy landscapes, magnetic forces, and magnetic-order-dependent energy-volume relations within the same general framework. For H-phase monolayer VSe\(_2\), stochastic spin-dynamics simulations using the learned magnetic effective fields locate the finite-size magnetic ordering crossover at 415–435~K, in close numerical agreement with the reported experimental value of \(418.5\pm7.8\)~K. These results establish Cartesian tensor message passing as a general route for connecting first-principles magnetic energetics with large-scale atomistic simulations of coupled structural and spin phenomena.
Read the full abstract
Machine Learning
Coarse composition suffices: tabular in-context learning for multi-activity antimicrobial peptide profiling
Antimicrobial peptides (AMPs) often act against multiple pathogen classes, making multi-label activity prediction a more realistic screening target than binary antimicrobial classification. The ESCAPE benchmark formalizes this setting, but leading approaches typically rely on multimodal, structure-conditioned deep models that are costly to train and tune. We show that a simple, sequence-only pipeline can match and surpass these methods by combining 330 interpretable sequence descriptors with TabPFN, a tabular foundation model that performs in-context prediction in a single forward pass without gradient-based training or hyperparameter search. On ESCAPE (82,359 peptides; five labels), a label-powerset TabPFN model achieves mAP-5 = 77.8%, improving on the previously best reported 72.1%. A probabilistic classifier chain is the first method to match or exceed the best published average precision on each of the five labels simultaneously. The gains persist under the prior state-of-the-art single-fold training protocol, indicating they are not a training-set-size artefact, and are largest for remote homologues (+11.2 points below 30% sequence identity). Ablations further show that predicted structure is unnecessary at inference and that performance is not driven by any single descriptor family: ten global physicochemical scalars recover 91% of full-feature performance. Finally, explicitly modelling label dependence yields targeted benefits for scarce activities and supports ranking which activity to assay next from partial positive evidence.
Read the full abstract
Combinatorics
An Improvement to the Upper Bound for Marton's Covering Conjecture
Marton's covering conjecture studies finite sets in high-dimensional binary spaces whose pairwise sums create relatively few new elements. It predicts that every such set can be described efficiently by shifted copies of one linear subspace of comparable size. Gowers, Green, Manners, and Tao [GGMT25] proved the conjecture with exponent $12$. Liao [L24] improved the exponent to $9$. We improve it further to $5.287$.
Read the full abstract
Strongly Correlated Electrons
Holographic Representations of Topological Quantum Criticality: Emergent Symmetry Approach around the Bott Clock
In this article, we apply the idea of emergent symmetries to construct holographic representations of a broad class of topological quantum critical points (tQCPs) that appear in fermionic classes in the Bott clock. We show explicitly that around the Bott clock, an emergent symmetry $U_{EM}$ exists at a tQCP between different symmetry protected gapped phases with protecting symmetry $G_p$ in $d$ spatial dimensions. This emergent symmetry can be used to construct a $(d+1)$ spatial dimensional lattice model with a properly upgraded symmetry such as $U_{EM}\times G_p$ or $U_{EM} \rtimes G_p$, etc. By doubling the degrees of freedom of the adjacent topological class in $d$-dimensions, we successfully show that the $(d+1)$-dimensional lattice models constructed in this way have the desired enlarged symmetry groups that precisely belong to the adjacent topological classes, counterclockwise around the Bott clock. Furthermore, $d$-dimensional boundaries of these $(d+1)$ dimensional lattice models of gapped topological classes are shown to exhibit identical infrared dynamics as the corresponding tQCPs in the $d$-dimensional adjacent topological phases in the Bott clock, with lower symmetries.
Read the full abstract
Statistics Theory
Unconditional Finiteness of the Number of Critical Points of Gaussian Mixture Densities
By slightly adapting Wang's recent important argument for the homoscedastic setting, we prove that every finite heteroscedastic Gaussian mixture has only finitely many critical points, and, consequently, only finitely many modes.
Information Retrieval
Beyond Ranking Accuracy: Evaluating LLM-Cited Feature Rationales for Next Basket Repurchase Recommendation
Next-basket repurchase recommendation is commonly formulated as a ranking task: given a customer's purchase history, the system ranks previously purchased items that may be needed again. In production settings, however, ranking accuracy is only one component of recommendation quality. Customers may also benefit from concise evidence about why an item is recommended now. Large language models (LLMs) offer a potential way to surface such evidence through feature-based, human-readable rationales grounded in interpretable behavioral signals. We construct repurchase features spanning cadence, frequency, recency, user behavior, and item popularity, and evaluate LLMs on two public grocery datasets and one proprietary retail dataset. We investigate (1) whether off-the-shelf LLMs can use these features as next-basket scorers relative to heuristic and supervised rankers, and (2) whether LLM-cited features carry outcome-grounded ranking signal. For the latter, we compare LLM-cited features with model-specific attribution methods under a cross-model feature-masking protocol that measures ranking degradation after masking selected features. Our results show that LLM scores are not competitive with supervised rankers, suggesting that off-the-shelf LLMs should not be used as standalone repurchase recommenders. However, changes in prompt and evidence representation can improve outcome-grounded feature-masking results in some settings even when ranking performance does not improve; the effect is dataset-dependent and does not consistently match attribution baselines. These findings suggest a practical role for LLMs as validated explanation components rather than primary rankers, with rationale quality evaluated separately from ranking accuracy.
Read the full abstract
Cryptography and Security
A Roadmap to Available ICS Datasets and Testbeds for Cybersecurity Research
Industrial Control Systems (ICS) are the backbone of many critical infrastructure sectors; however, their growing level of connectivity, long lifespan and integration with the Information Technology (IT) environment introduces numerous cybersecurity challenges. The merging of Operational Technology (OT) and IT along with the deployment of Industry 4.0 technologies increases the attack surface of ICS environments, which in turn makes them more vulnerable to advanced cyber threats. Therefore, many researchers have shown interest in the field of cybersecurity of ICS. The topics of intrusion detection, anomaly detection, threat intelligence, attack simulation and resilience assessment of ICS have received much attention. Nevertheless, the development and testing of cybersecurity solutions for ICS remains to be challenging due to the lack of appropriate datasets and experimental environment. The main objective of this paper is to provide the roadmap of existing ICS cybersecurity datasets, testbeds and digital twins. This paper presents various taxonomies along with systematic analysis of architecture, characteristics, capabilities, pros and cons of these tools. The results of the analysis demonstrate the presence of persistent problems such as lack of standardized benchmarking datasets, lack of modern attack scenarios, insufficient number of datasets based on real operational traffic and difficulty in validating artificial intelligence-driven cybersecurity solutions. In addition to summarizing current research on ICS cybersecurity datasets and testbeds, this roadmap provides the identification of research gaps and recommendations on creation of new tools.
Read the full abstract
Statistics Theory
Bergsma–Dassios Sign Covariance Characterises Independence for Arbitrary Real-Valued Bivariate Laws
Bergsma–Dassios sign covariance $\tau^*$ is a rank-based population measure of dependence. Building on zero-characterisation results under specific regularity regimes, we prove that $\tau^*(X,Y)=0$ characterises independence for every real-valued bivariate distribution, including mixed and singular laws. For the unnormalised four-sample convention for $\tau^*$ defined in Subsection 4.3 and the unscaled Blum–Kiefer–Rosenblatt functional $\mathscr {B}$, the proof gives the quantitative inequality $\tau^*\ge 2\mathscr {B}$. This is a population identification result; no new sample-level limit theorem is claimed. The argument first encodes finite ordered distributions with rational cell probabilities by labelled path trees and applies a nonnegative sum-of-squares representation for a quartet covariance. Rational approximation and nested quantisation then remove all support and regularity restrictions. On finite uniformly weighted label sets, the tree framework also relates an edge-weighted quartet quantity to empirical distance covariance squared. As a separate combinatorial consequence, it yields the asymptotic $2/3$ upper bound for the quartet distance between binary phylogenetic trees.
Read the full abstract
Combinatorics
Vertex-transitive strongly regular graphs in the switching class of doubly transitive two-graphs
Let $G$ be a permutation group that acts $2$-transitively on the finite set $V$ and let $\mathcal{T}=(V,T)$ be a two-graph whose automorphism group contains $G$. In this paper, we classify those strongly regular graphs $\Gamma$ with vertex set $V$ whose automorphism group is a transitive maximal subgroup of $G$ and whose associated two-graph is $\mathcal{T}$. In doing so, we obtain a new family of vertex-transitive strongly regular graphs whose associated two-graph arises from $P\Sigma L(2,q)$.
Read the full abstract
Sound
Ouroboros: Self-Referential Backdoor Attacks on Speech Enhancement via Clean Audio Triggers
Speech enhancement models are widely deployed as frontend modules in real-time speech services, yet their vulnerability to backdoor attacks remains unexplored. Existing backdoor methods are confined to classification tasks and rely on active trigger injection, an assumption incompatible with the passive processing nature of speech enhancement models. In this paper, we propose Ouroboros, a novel backdoor attack framework that leverages the ideal clean outputs of speech enhancement models as natural triggers, enabling inference-time activation without any external trigger injection. Extensive evaluations show Ouroboros achieves near-perfect attack success rates with minimal performance degradation on diverse models and datasets. Physical-world validations confirm that naturally recorded, unaltered clean audio can reliably activate the backdoor. Moreover, Ouroboros generalizes to targeted content-tampering attacks and remains effective against common filtering and finetuning defenses.
Read the full abstract
Machine Learning
Learning PDE Time-Stepping with Neural Cellular Automata
Classical numerical solvers for partial differential equations (PDEs) are computationally expensive to solve repeatedly across varying initial conditions, motivating the need for learned surrogates. In this paper, we propose a trainable Neural Cellular Automata (NCA) based surrogate model for learning long time PDE dynamics. Rather than mapping an entire initial field to a full trajectory in one shot, our proposed model learns a small, local, homogeneous update rule that is applied identically and repeatedly at every grid cell, mirroring the locality of differential operators. We benchmark this framework against three baselines: PDE - Net, a modified physics-informed neural network (PINN), and a Fourier Neural Operator (FNO), on five canonical PDEs (heat, advection, Burgers, Allen - Cahn, and Fisher - KPP), evaluated at temporal domain two times beyond the training temporal domain. The proposed model achieves the lowest long-horizon relative errors on the majority of the experiments.
Read the full abstract
Computation and Language
Do Small Models Use the Law You Give Them? Measuring Context Use on a Bilingual Bangladesh Legal Benchmark
Fine-tuning can improve legal question-answering accuracy without improving how models use law supplied in context. We study this distinction in bilingual Bangladeshi legal QA, where observed errors can arise from answer scoring, retrieval, or failure to use relevant law. We construct a hierarchy-preserving statutory corpus, 2,165 reviewed bilingual fine-tuning examples, and a 150-item supplied-law control. We evaluate six instruction-tuned models: Llama-3.2-1B, Llama-3.2-3B, Qwen3.5-0.8B, Qwen3.5-2B, Qwen3.5-4B, and Gemma-4-E2B, with three LoRA seeds per model. To separate effects, we combine constrained option-letter scoring, cyclic option rotation, and controlled removal of the governing provision. On 398 Bar Council outputs, an exact-line parser attributes an accuracy gain of 50.0% to the Qwen3.5-2B seed-42 adapter, whereas option scoring yields only $3.0\%$. For Gemma-4-E2B, the two scoring methods favor different systems. When the governing provision is guaranteed to be present, five of six reference models improve by $14.7\%-19.3\%$ under the four-order criterion. Removing that provision reduces accuracy by $8.0\%-15.3\%$ for models and by $13.8\%-14.9\%$ points for their adapters. However, difference-in differences estimates show no increase in reliance on the governing provision after fine-tuning. Results show that legal adaptation claims require separating scorer, retriever, and model effects. Our Code and data are available at https://anonymous.4open.science/r/bangladesh-legal-qa-11E3
Read the full abstract
Sound
Parallel Time-Band Mixing with Learned Observation-Adding for Robust ASR Front-Ends
Speech enhancement is often used as a front-end for robust ASR, yet recurrent temporal and cross-band modules introduce sequential dependencies that reduce parallel efficiency. In this paper, we present a sequence-parallel band-split enhancement front-end built on a Parallel Time-Band Mixer (PTBM) block that eliminates within-block recurrent unrolling. PTBM integrates intra-band temporal mixing and per-frame cross-band attention within a unified parallel architecture, enabling efficient contextual modeling across both time and frequency dimensions. The system retains the mask-plus-residual reconstruction interface and introduces learned Observation-Adding (LOA) to suppress ASR-sensitive artifacts without development-set tuning. Experiments on DNS Challenge and CHiME-4 with frozen Whisper back-ends show that the proposed front-end consistently reduces word error rate relative to recurrent band-split baselines while requiring only 0.96 M parameters and 0.58 GMAC/s for the front-end network.
Read the full abstract
Computation and Language
Sequential Trajectories and Simultaneous Blending: Multi-Emotion Modeling for Instruction-Following TTS
Natural-language instructions enable flexible control of synthesized speech, yet emotional TTS systems primarily model a single utterance-level affect, leaving multi-emotion control underexplored. We study two complementary multi-emotion TTS tasks: emotion trajectory, which spans several ordered affective stages, and emotion blending, in which multiple emotions coexist throughout an utterance. These tasks expose a supervision mismatch: supervised fine-tuning (SFT) does not explicitly evaluate emotion features, while single-emotion rewards provide neither structure-aware feedback for trajectory completion nor pair-aware feedback for blending. We introduce HybridEmo, a post-training framework that initializes both tasks with SFT and then aligns the speech-token policy through Group Relative Policy Optimization using a sample-aware hybrid reward. For trajectory samples, segment-aligned consistency combines average and weakest-stage evidence to preserve the correctness and completeness of prescribed stages. For blending samples, a GMM-based reward combines frame-level support from the union of target-emotion anchors in an offline emotion space with an utterance-level weaker-target margin. Both branches share an ASR reward and are routed within a unified policy. On MultiEmo-Test, HybridEmo significantly improves trajectory correctness and blending intensity, without a noticeable degradation in speaker similarity. Human evaluation prefers HybridEmo to CosyVoice 3 and EmoVoice-0.5B, with nearly balanced preferences against Qwen3-TTS.
Read the full abstract
Mesoscale and Nanoscale Physics
Scalar Spin Chirality from Dissipative Pumping and Lamb-Shift Precession
How electrical driving generates and reverses scalar spin chirality in an interacting system with a real Hamiltonian and no magnetic flux remains a fundamental open question. We show that dissipative tunneling prepares an orbital pseudospin, while virtual charge fluctuations generate a noncollinear Lamb field that rotates it toward the chiral direction. Bias changes their relative orientation and thereby reverses the current. This reservoir-induced orbital Hanle effect converts a nonchiral orbital polarization into scalar spin chirality. Within the ground-state space, the circulating current is determined by the scalar spin chirality. Hierarchical-equations-of-motion calculations show that the low-bias reversal survives beyond the second-order weak-coupling description, while the quadratic growth of chirality after a sudden voltage switch identifies precession. More broadly, time-reversal-odd responses can emerge in open quantum systems from the interplay between dissipative state preparation and coherent precession.
Read the full abstract
Machine Learning
Generative multi-domain transfer learning for fault detection in data-scarce wind turbines
Normal behavior models have shown promise for reliable fault detection in wind turbines. However, these unsupervised anomaly detection models require sufficient fault-free training data to learn the normal operation behavior of turbines. Under data scarcity, for example in newly deployed wind turbines, these models may result in poor fault detection performance. In this work, we propose a multi-domain generative domain mapping approach based on Star Generative Adversarial Networks (StarGAN) to improve fault detection on data-scarce wind turbines. Our model maps SCADA measurements from a data-scarce turbine to resemble those of several data-rich turbines. By preserving the operational state during translation, faults occurring in a data-scarce domain can be mapped and detected by reliable pre-trained normal behavior models of data-rich domains. Highlighting the benefits of an ensemble fusion strategy, we show that under severe data scarcity our method can produce anomaly scores comparable to models trained on large representative datasets. Our approach can consistently outperform models trained on scarce data when less than 2 weeks of training data are available. With just 2 weeks of accumulated training data, we achieve an anomaly score similarity that is, on average, +16% higher than conventional fine-tuning, and +10% higher than single-source domain mapping. As a step towards unsupervised model selection, we propose a proxy metric that detects poor performance at training time, despite an absence of anomalies. Our study presents the potential and challenges of multi-domain mapping for wind turbine fault detection under unrepresentative training data.
Read the full abstract
Artificial Intelligence
Ignorance or Incompetence? Constructing Knowledge-Gated, Verifiable Tasks for LLM Agents
Professional agent tasks often depend on conventions that are absent from public corpora, yet benchmarks rarely control whether an agent has access to those conventions. We introduce a knowledge-gated task-construction protocol that separates a task instruction from a compact artefact containing private conventions, reference tables, and utility operators. Construction-time provenance, byte-identical task instructions across the provided- and withheld-artefact conditions, leak audits, and executable witnesses make dependence on the artefact explicit and testable. Across fifteen calibration tasks, one frontier agent configuration achieves a 68.0% pass rate with the artefact and 0% without it; on one task, a plausible but incorrect artefact also yields 0% across five trials. Deterministic solvers and rule corpora provide exact ground truth for structured tasks, while named criterion-level rubrics support outputs that cannot be checked by a single executable oracle. A configuration-relative calibration screen retains seven tasks satisfying our five-trial empirical knowledge-gating screen. These experiments validate the behavior of the construction protocol; they do not establish that the retained tasks improve post-training. We publicly release part of the task suite and supporting tooling at https://github.com/DatagridsAI/Knowledge-Gated-Task-Construction.
Read the full abstract
Mathematical Finance
Optimal Block Time for AMM Liquidity Providers under Jump-Diffusion Prices
Loss-versus-Rebalancing (LVR) is the dominant adverse-selection cost borne by liquidity providers on automated market makers. Under geometric Brownian motion, arbitrage profit scales with the probability of a profitable block, which vanishes as the block time $\Delta t \to 0$; this is the standing argument for ever-shorter blocks. Modeling the reference price instead as a jump-diffusion, I show that the constant-product LVR rate splits into a diffusion channel carrying the known multiplier $F(\gamma/(\sigma\sqrt{\Delta t}))$ and a jump channel $\lambda V \cdot G(\gamma;m,\delta^2)$ carrying no $\Delta t$, the two interacting only through an explicitly bounded remainder. The block schedule therefore governs only one channel. For symmetric jump laws the jump channel is moreover an exact lower bound, $\ell(\Delta t) \ge \lambda V G > 0$, so the rate does not vanish as $\Delta t \to 0$, and descends slowly, as $\sqrt{\Delta t}$. At Ethereum's calibrated 12-second slot the rate is 471 bp/yr against a floor of 125, so only three quarters of LP loss is schedule-addressable. At Solana's 400 ms slot the jump channel already dominates. Netting the rate against per-block consensus cost, the LP-side optimal block time is invariant in pool size and in every jump parameter $(\lambda,m,\delta)$: jumps shift the level of LP loss but not the planner's marginal tradeoff. Volatility, the fee tier, and consensus cost set the optimum, near 8 s. However, LVR is only one input to block-time welfare, so this bounds the LP-side contribution rather than settling the design question.
Read the full abstract
Computation and Language
On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability
We describe the architecture and ablations of Qwen3.8-Flash-Next, a sparse mixture-of-experts model with 125B parameters, 6B activated per token, and additional 51B parameters of n-gram embedding tables held off the accelerator. On fourteen pre-training benchmarks the model leads the 397B-A17B predecessor on eight and trails it on the rest by at most 2.6 points, at 1/3 the activated parameters, 1/3 the training tokens, and roughly 1/9 the training FLOPs. Token mixing uses a layer-wise hybrid of Gated DeltaNet (GDN) and global attention, with one full-attention layer in every four; at continued-pretraining time those full-attention layers are replaced by Qwen Sparse Attention (QSA), which scores context at micro-block granularity with a compressed lightweight indexer. The residual stream is widened to four branches and read through an elementwise gate, a design we call the Gated Residual (GR). Capacity is added outside the backbone by a single n-gram embedding layer whose tables are prefetched from host memory. We evaluate every candidate change along three axes: loss together with downstream benchmarks; the cost of the change in training, prefill and decode; and its effect on the optimal hyperparameters and training stability. Loss and downstream accuracy do not always move together: enlarging the n-gram vocabulary lowers loss monotonically while downstream accuracy saturates. The architecture and the Muon optimizer together shift the optimal learning rate and batch size upwards, render batch-size warmup unnecessary, and substantially improve stability under stress tests. Loss, benchmarks, efficiency and stability form one design problem. Solved jointly, they yield a recipe that is simultaneously more efficient, more capable and more stable.
Read the full abstract
Computation and Language
Beyond Token-Level Guidance: Inference-Time Alignment of Specialized LLMs via Cross-Family Representation Steering
Large language models (LLMs) finetuned for specialized domains represent crucial high-impact applications. Inference-time alignment improves safety degraded from specialization finetuning without requiring substantial computational resources, complementing finetuning-based methods with an easy-to-use, plug-and-play solution. However, existing inference-time methods fail to reliably improve safety without disrupting domain capability. We identify the root cause as complementary expertise orthogonality: specialized base models and general-domain guidance models have orthogonal competencies, making the guidance signal unreliable for specialized generation. This primarily manifests as stop token interference, where the guidance model's tendency toward continuation overrides the base model's decision to stop, burying correct answers under guidance-induced continuation. To address this problem, we propose CREST, an inference-time alignment method that steers base model hidden representations using safety directions extracted from a guidance model of any family, avoiding token-level structural limitations entirely. CREST improves safety where specialization has weakened it while preserving both domain-specific capability and the safety of already well-aligned models, outperforming baselines by up to 22.2% on safety benchmarks. Our code is available at: https://github.com/DecayingSeart/CREST.
Read the full abstract
Systems and Control
Converse Barrier Certificates for Set-Based Stochastic Reach-Avoid Verification
Recent work established sufficient and necessary barrier-like conditions for infinite-horizon reach-avoid verification of stochastic discrete-time systems from a single initial state. Whether such a converse characterization extends to a set of initial states, however, remains open. In this paper, we answer this question affirmatively for compact initial sets. We consider a uniform reach-avoid specification requiring the reach-avoid probability to exceed a prescribed threshold for every initial state in a compact set. Under appropriate assumptions, including continuous system transitions, together with a strict uniform probability margin, we extend the pointwise converse characterization to the uniform setting.
Read the full abstract
Machine Learning
Online Estimation of Dynamic Origin-Destination Matrices Using Reinforcement Learning with Link-Flow Propagation Guidance
Dynamic origin-destination (OD) matrix estimation calibrates time-dependent input demand for simulations to reproduce observed link flows. Reinforcement learning is well suited to online estimation because a trained policy estimates demand in a single evaluation, but varying target flows complicate learning from aggregate rewards. We propose reinforcement learning with link-flow propagation guidance (LFPG-RL), combining simulated vehicle propagation records with downstream errors to guide individual OD components. Using 250 weekday link-flow trajectories from a Melbourne arterial network, LFPG-RL achieves a mean test root mean squared error (RMSE) of 7.41 vehicles per 15-minute interval, mean absolute percentage error of 31.24%, and Pearson correlation of 0.987. Comparisons with optimisation, filtering and reinforcement learning without guidance show a 40.1% RMSE reduction over the strongest baseline, gradient descent with propagation guidance. LFPG-RL enables rapid demand updates that keep traffic simulations consistent with observed conditions, supporting the evaluation of traffic management strategies.
Read the full abstract
Computer Vision
Knowing Beyond the Known: Reinforced Knowledge Specification for Multi-Label Class-Incremental Learning
Existing class-incremental learning methods struggle in multi-label scenarios (MLCIL) due to the inherent contradiction of learning objectives arising from co-occurring and incomplete labels. We argue that the core obstacle is the model's ambiguous boundary between known and unknown knowledge, which undermines historical knowledge retention, complicates current task learning, and limits adaptability to future concepts. To address this, we propose KBK (Knowing Beyond the Known), a reinforced knowledge specification framework that explicitly models what is known or not to unify historical, current, and prospective learning. Specifically, to clarify known knowledge, we develop a hierarchical feature purification module that disentangles fine-grained class-specific features from global features, where high-level semantic abstraction is reinforced with low-level visual features. Additionally, an uncertainty-aware recall enhancement strategy suppresses unreliable predictions based on distribution priors, improving the quality of historical recall. For probing the unknown, KBK leverages semantic correlations to synthesize informative unknown features under co-occurring, preserving embedding space for future learning. Furthermore, to mitigate heterogeneous forgetting, we design a category-balanced gradient compensation loss that dynamically reweights gradient backpropagation according to forgetting speeds. Experiments on multiple benchmarks validate the effectiveness and robustness of KBK, which surpasses prior best methods by 2.7% in Avg. Acc on MS-COCO B0-C10 setting even without any replay buffers.
Read the full abstract
Machine Learning
Context Staircase: Signature-Aligned Dynamics of Token Embeddings under Small Initialization
Token embeddings are the basic representational units that connect discrete tokens with continuous computation in language models. Although modern language models learn embeddings from random initialization through gradient-based training, the dynamical mechanism by which meaningful embedding structures emerge remains unclear. In this work, we identify that the evolving embedding structures are closely related to token-conditioned label and contextual distributions, which we formalize as probability signatures. We observe a progressive learning process, which we term Context Staircase: embeddings learn the low-order statistic signatures of the data before the high-order ones. More specifically, we observe that early in training they align with the simplest, context-free signature linking a token to its label, and as training proceeds, they progressively reflect signatures involving more and more context tokens. We then analyze the gradient flow of embeddings under small initialization to explain this phenomenon, deriving embedding evolution equations for feed-forward and self-attention architectures. We further extend these observations to real language-model training. Finally, we show that these embedding structures play an important role in both task learning and the incorporation of semantic structure into the embedding space. Overall, our results provide a dynamic explanation of how data statistics and architecture jointly shape token embeddings in language models, and reveal an implicit bias in the space of data statistics: training proceeds from simpler, low-order statistical relations toward increasingly complex, context-dependent ones.
Read the full abstract
Audio and Speech Processing
Weakly Supervised Tabla Stroke Transcription via an Adaptive Dynamic Rhythm Language Model (ADRM)
Tabla Stroke Transcription (TST) is central to the analysis of rhythmic structure in Hindustani music, yet it remains challenging due to complex and dynamic rhythmic organization and the scarcity of strongly annotated data. Existing approaches largely rely on fully supervised learning with onset-level annotations, which are costly and impractical at scale. This work addresses TST in a weakly supervised setting, using only symbolic stroke sequences without temporal alignment of onsets. We propose a framework that combines a Connectionist Temporal Classification (CTC)-based acoustic model with a sequence-level rhythmic language model for rescoring, similar to that used in automatic speech recognition. The acoustic model produces a decoding lattice, which is refined using an Adaptive Dynamic Rhythm Language Model (ADRM) that combines $t\bar{a}la$-conditioned symbolic rhythmic regularities with local stroke dynamics. Moreover, we release a new performance-recorded tabla dataset, named Tabla Improvisation Dataset, along with a complementary synthetic dataset for sequence-level weakly supervised TST. Experiments demonstrate consistent and substantial reductions in stroke error rates with ADRM compared to those with acoustic-only decoding, confirming the benefit of incorporating symbolic rhythmic regularities during lattice rescoring for accurate transcription.
Read the full abstract
Quantum Physics
Quantum-interference metrology of dissipative Kerr solitons
Dissipative Kerr solitons in optical microresonators underpin chip-scale frequency combs with applications ranging from coherent telecommunications to precision spectroscopy. Yet the characterization of their intrinsic femtosecond temporal structure remains challenging, as the low pulse energy and broad spectral bandwidth necessitate optical amplification and careful dispersion compensation in conventional ultrafast diagnostics, both of which can significantly distort the waveform. Here we demonstrate a quantum-interference metrology of microcomb solitons based on Hong-Ou-Mandel interference. By attenuating the soliton stream to the single-photon level and measuring fourth-order interference, we directly retrieve near transform-limited pulse durations without amplification or dispersion management, remaining accurate even after propagation through 25 km of standard fiber. The same interferogram also provides direct access to the temporal separations in multi-soliton states by converting inter-soliton separations into additional interference dips at corresponding delays, enabling sub-picosecond characterization of their intracavity temporal structure. This quantum-inspired paradigm introduces a fundamentally new metrological approach that is immune to amplification and dispersion distortions, offering a powerful tool for the characterization of complex soliton physics.
Read the full abstract
Astrophysics of Galaxies
Testing SALT Approximations with Numerical Radiative Transfer Code. II. Thermal and Microturbulent Line Broadening
Forward models that connect galactic winds to their predicted spectral-line profiles have proved effective for inferring wind properties in controlled settings, but important limitations remain. In particular, many models rely on the Sobolev approximation to solve the radiative transfer equation and neglect line broadening caused by thermal and turbulent motions within the wind. In the first paper of this series, we demonstrated that neglecting this broadening in Semi-Analytical Line Transfer (SALT) models can bias the recovery of fundamental wind properties from mock observations. Here, we extend the SALT framework to incorporate this motion by solving the radiative transfer equation in the single-scattering limit. We treat re-emission using an escape-probability approach similar to that adopted under the Sobolev approximation, while allowing photons to escape from resonance regions of finite thickness. We validate the model and investigate parameter degeneracies by fitting mock spectra generated with Monte Carlo radiative transfer simulations assuming identical outflow configurations. We identify a degeneracy between the Doppler-broadening parameter and the radial density and velocity profiles: shallower density and velocity gradients can mimic the effects of greater velocity dispersion. Nevertheless, integrated quantities are well recovered. Over the range $13 \leq \log(N_{\mathrm{Si}^+}/\mathrm{cm}^{-2}) \leq 18$, the recovered ionic column densities have a scatter of 0.26 dex and are systematically overestimated by 0.22 dex. Mass-outflow rates evaluated at the terminal wind radius have a scatter of 0.88 dex and are systematically overestimated by 0.51 dex. These results represent substantial improvements over previous versions of the model.