Academic Intelligence · Curated Daily

Explore the Frontier of Global Academia

AcademicHub aggregates real-time literature from top journals and preprint platforms. Build your personal research radar and let large language models compile cross-disciplinary analysis briefings automatically.

01.
medRxiv (Medicine) 2026-06-22

Age-related changes in acoustic cue use for speech-in-speech perception

Authors:

Acoustic cues such as pitch and spatial location allow listeners to attend to a target speaker and ignore competing talkers, aiding speech recognition in background noise. Diminished ability to utilize acoustic cues for speech stream segregation may thus contribute to older adults' challenges hearing in noise. Adults aged 18-74 completed a speech-in-speech identification task with three conditions containing 1) only pitch cues (fundamental frequency), 2) only spatial cues (interaural time differences; ITDs), and 3) both pitch and spatial cues for segregating a target talker from competing talkers. Hearing thresholds at standard and extended high frequencies (EHFs), auditory brainstem responses (ABRs), and digit span scores were acquired to examine the influence of sensory and cognitive factors on use of each acoustic cue for speech-in-speech recognition. Significant differences were observed between cue condition scores indicating that use of the available cue(s) drove performance. ABR metrics were not a significant predictor but digit span scores significantly predicted scores on all three cue conditions. Working memory abilities therefore set a baseline for participants' speech-in-speech recognition regardless of the acoustic content. Hearing thresholds at standard frequencies significantly predicted scores on the Pitch condition. EHF hearing thresholds better predicted Spatial and Both Cue condition performance, suggesting that EHF thresholds represent auditory processing important for coding ITDs. Age group analysis revealed that older adults (aged 40+) performed significantly more poorly on all cue conditions of the speech-in-speech recognition task relative to younger adults. Age-related changes in auditory sensory processing may therefore impair older adults' speech-in-noise perception by reducing their ability to use acoustic cues for segregating target and competing speech.

02.
arXiv (CS.LG) 2026-06-24

A Fast and Effective Method for Euclidean Anticlustering: The Assignment-Based-Anticlustering Algorithm

arXiv:2601.06351v2 Announce Type: replace Abstract: Anticlustering is an NP-hard combinatorial optimization problem that consists of partitioning a set of objects into equal-sized groups called anticlusters such that the objects in the same anticluster are as dissimilar as possible and thereby representative of the entire set of objects. Here we study the case where the dissimilarity metric is the squared Euclidean distance between the respective feature vectors. Applications of Euclidean anticlustering include social studies, cross-validation, creating mini-batches for stochastic gradient descent, and finding balanced K-cut partitions. In particular, machine-learning applications such as mini-batch generation involve million-scale datasets and very large values of K, making scalable anticlustering algorithms essential. We propose a new algorithm, the Assignment-Based Anticlustering (ABA) algorithm, that scales to instances with millions of objects and hundreds of thousands of anticlusters within seconds to minutes, which is far beyond what existing anticlustering methods can manage. We demonstrate here, via an extensive computational study, that our algorithm outperforms existing anticlustering methods in both solution quality and running time. This is so also for anticlustering with categories. For the related problem of balanced K-cut partitioning, our algorithm is superior to the well-known METIS method. The code of our algorithm is available on GitHub.

03.
arXiv (quant-ph) 2026-06-11

Time-Frequency Grid States for Reconstruction and Correction of Channel-Induced Distortion in Entangled Photons

arXiv:2606.12216v1 Announce Type: new Abstract: Characterization of time-frequency (TF) quantum states requires reliable reconstruction of their TF distributions. However, imperfect transmission or measurement channels can distort reconstructed joint spectral intensities (JSIs), especially when the underlying perturbation mechanism is unknown. Here, we experimentally demonstrate a reconstruction and correction framework that uses a TF grid state as an intrinsic frequency-domain reference. By analyzing the displacement of the grid points, a Gaussian process regression model is employed to reconstruct a correction mapping for the nonlinear coordinate deformation without assuming a prior physical model of the distortion. The learned mapping reduces the residual coordinate deviation of the TF grid state by approximately a factor of 11 and, when applied to an independent frequency-entangled test state, improves the Gaussian-shape fidelity from 76.2\% to 90.0\%. These results establish TF grid states as practical metrological resources for diagnosing and correcting distortions in TF quantum systems, providing a pathway toward distortion-resilient quantum communication and high-dimensional quantum information processing.

04.
arXiv (CS.AI) 2026-06-11

The Impossibility of Eliciting Latent Knowledge

arXiv:2606.12268v1 Announce Type: new Abstract: Advanced AI systems have extensive knowledge of their environments; in fact, their knowledge may (far) exceed that of their developers or users. Consequently, a desirable property for an AI system is that it is honest – that it accurately reports its beliefs about the world. Designing an AI system to be honest may be difficult, especially if we want to ask it questions about latent variables in the environment – variables which are hidden from the human interacting with it. This gives rise to the problem of eliciting latent knowledge (ELK): the problem of training an AI agent to honestly report its beliefs. In this paper, we make ELK formally precise using Causal Influence Diagrams (CIDs). CIDs can be used to describe the relationship between an agent's training environment and its subjective representation of the world. We use CIDs to formalise the distinction between observable and latent variables, to specify what exactly it means for an agent to be honest, and to formally define goal misgeneralisation. We show that, under certain circumstances, developers can incentivise an agent to honestly answer questions by providing correct feedback during training. However, a natural, but undesirable, way for an agent to generalise is to provide answers which humans would evaluate as true, rather than honest answers. We prove an impossibility theorem stating: There is no feedback-based training strategy that depends only on agent behaviour and with certainty produces an honest agent, even if feedback is perfect during training.

05.
arXiv (CS.CL) 2026-06-16

Evaluative Judgement in Teaching AI-based Translation: A Class-room Case Study of AI-Mediated Translation and Post-Editing

Authors:

Drawing on 23 anonymized student pro-jects from a fourth-year Machine Transla-tion and Post-editing course in a BA-level translation programme, this paper exam-ines how structured comparison of gen-eral-purpose LLMs and online MT sys-tems can elicit evaluative judgement in AI-mediated translation. Students translat-ed short specialised English Wikipedia texts into Catalan or Spanish, generated four system outputs, evaluated them using automatic metrics and human adequa-cy/fluency assessment, selected one output for post-editing, and justified their deci-sion in written reports. Descriptive counts are reported for all 23 projects, while qualitative interpretation is based on the 22 cases accompanied by written reports. Results show that students did not treat automatic metrics as final authority: final post-editing selections often diverged from metric rankings and were justified through adequacy, fluency, terminology, naturalness, and expected post-editing ef-fort. The study therefore does not bench-mark systems under controlled conditions; it analyses how students justified system choice within an authentic classroom as-signment.

06.
arXiv (CS.AI) 2026-06-16

When Do We Need LLMs? A Diagnostic for Language-Driven Bandits

arXiv:2604.05859v2 Announce Type: replace Abstract: We study Contextual Multi-Armed Bandits (CMABs) for non-episodic decision-making problems where the context includes both textual and numerical information (e.g., recommendation systems, dynamic portfolio adjustments, offer selection; all frequent problems in finance). While Large Language Models (LLMs) are increasingly applied to these settings, utilizing LLMs for reasoning at every decision step is computationally expensive, and uncertainty estimates are difficult to obtain. To address this, we introduce LLMP-UCB, a bandit algorithm that derives uncertainty estimates from LLMs via repeated inference. However, our experiments demonstrate that lightweight numerical bandits operating on text embeddings (dense or Matryoshka) match or exceed the accuracy of LLM-based solutions at a fraction of their cost. We further show that embedding dimensionality is a practical lever on the exploration-exploitation balance, enabling cost-performance tradeoffs without prompt complexity. Finally, to guide practitioners, we propose a geometric diagnostic based on the arms' embeddings to decide when to use LLM-driven reasoning versus a lightweight numerical bandit. Our results provide a principled deployment framework for cost-effective, uncertainty-aware decision systems with broad applicability across AI use cases.

07.
arXiv (CS.CV) 2026-06-12

Radar-Guided Polynomial Fitting for Metric Depth Estimation

We propose POLAR, a novel radar-guided depth estimation method that introduces polynomial fitting to efficiently transform scaleless depth predictions from pretrained monocular depth estimation (MDE) models into metric depth maps. Unlike existing approaches that rely on complex architectures or expensive sensors, our method is grounded in a fundamental insight: although MDE models often infer reasonable local depth structure within each object or local region, they may misalign these regions relative to one another, making a linear scale and shift (affine) transformation insufficient given three or more of these regions. To address this limitation, we use polynomial coefficients predicted from cheap, ubiquitous radar data to adaptively adjust predictions non-uniformly across depth ranges. In this way, POLAR generalizes beyond affine transformations and is able to correct such misalignments by introducing inflection points. Importantly, our polynomial fitting framework preserves structural consistency through a novel training objective that enforces local monotonicity via first-derivative regularization. POLAR achieves state-of-the-art performance across three datasets, outperforming existing methods by an average of 24.9% in MAE and 33.2% in RMSE, while also achieving state-of-the-art efficiency in terms of latency and computational cost.

08.
arXiv (CS.AI) 2026-06-11

Are Frontier LLMs Ready for Cybersecurity? Evidence for Vertical Foundation Models from Dual-Mode Vulnerability Benchmarks

arXiv:2605.23243v2 Announce Type: replace-cross Abstract: We evaluate whether frontier LLMs are ready for cybersecurity through a dual-mode benchmark: white-box function-level vulnerability detection (VulnLLM-R, across C/Java/Python) and black-box web application security testing (five production-style applications with 118 ground-truth vulnerabilities across 20+ CWE families, which we will open-source). We test six frontier models (GPT-5.4, Codex~5.3, Claude Opus~4.6, Sonnet~4.6, Gemini~3.1~Pro and Gemini~3~Flash) and two domain-specialized models across four testing paradigms. Our findings are sobering: (1)~every frontier model produces 10-50% false positive rates in white-box detection, systematically over-predicting vulnerabilities; (2)~in black-box testing, frontier models achieve only 4-8% ground-truth coverage, improving to just 10-19% even with external security tools (Playwright MCP, Burp Suite MCP); (3)~structured penetration-testing methodology encoded in domain-specialized agents raises per-family detection above 50%, demonstrating that methodology, not scale, is the primary lever; and (4)~a domain-specialized defense model achieves the highest precision (0.904) and lowest false positive rate (9.7%) among all models, on a single GPU. We identify the absence of structured security testing traces end-to-end request/response sequences, failure-heavy data, and multi-step attack chains as the fundamental training data bottleneck, and propose self-play security testing as a data generation strategy. Our results make the case for vertical foundation models purpose-built for cybersecurity.

09.
arXiv (CS.CL) 2026-06-18

Rethinking Reward Supervision: Rubric-Conditioned Self-Distillation

Post-training of reasoning language models is commonly driven by supervised distillation and reinforcement learning with verifiable rewards. Distillation often relies on chain-of-thought annotations that are expensive to obtain and may themselves be noisy, incomplete, or partially incorrect; even when the final solution is correct, an imperfect rationale can interfere with learning. Reinforcement learning with verified rewards, on the other hand, typically compresses evaluative feedback into a scalar signal, obscuring which aspects of a response should be improved. We propose Rubric-Conditioned Self-Distillation, a framework that incorporates rubrics as structured, fine-grained feedback for on-policy self-distillation. Our method conditions the teacher model on criterion-level rubrics and uses it to provide token-level guidance on the student's own sampled trajectories. This design avoids treating a single reference rationale as the sole supervision target. Instead, rubrics specify what a strong response should satisfy, enabling more fine-grained credit assignment over the reasoning process than scalar reward optimization. We instantiate this framework with a two-stage pipeline that first learns to generate task-specific rubrics and then trains a rubric-guided reasoner. We evaluate on a diverse suite of science reasoning benchmarks and results show that rubric-conditioned self-distillation effectively converts rubric-level criteria into token-level guidance over the reasoning process, surpassing GRPO by 1.0 points and OPSD by 0.9 points on average.

10.
arXiv (CS.CV) 2026-06-16

Learned Image Compression for Vision-Language-Action Models

Vision-language-action (VLA) models increasingly rely on high-frequency multi-camera observations, making visual communication a major bottleneck for real-time robotic control in bandwidth-constrained or distributed deployment settings. Existing image and video codecs, however, are designed to preserve generic visual fidelity rather than the control performance of downstream VLA policies. In this work, we introduce SPARC (SPatially Adaptive Rate Control), a learned image compression framework tailored for VLA-driven robots. Our key observation is that the importance of visual information varies substantially across both camera views and spatial regions within an image. Based on this observation, SPARC employs a lightweight temporal mask selector that adaptively allocates bitrate over latent representations according to task relevance while leveraging temporal context. We further introduce a tilted rate loss that stabilizes training by reducing the tendency of entropy-based objectives to over-suppress rare yet task-critical visual patterns. Experiments on diverse robotic benchmarks, including RoboCasa365, VLABench, and LIBERO, show that SPARC consistently achieves stronger control performance than conventional image/video codecs and recent learned compression methods under the same bitrate budget. We additionally demonstrate real-world deployment benefits in remote-control settings, where our method substantially improves the bitrate-success tradeoff.

11.
PLOS Medicine 2026-06-12

Placenta accreta spectrum in the 21st century: Challenging dogma and redefining disorder

Authors:

by Eric Jauniaux, Helena C. Bartels, Yalda Afshar Placenta accreta spectrum (PAS) is a serious pregnancy complication caused by abnormal placental attachment to the uterus. In this Perspective, Eric Jauniaux and colleagues discuss emerging evidence that challenges our long-held pathophysiological understanding of PAS, and argue that a critical reassessment of definition, diagnosis, and management is overdue. In this Perspective, Jonathan Evans and colleagues discuss why restricting access to joint replacement surgery based on BMI alone is not supported by evidence, and highlight how such rest rictions risk exacerbating stigma, inequity and avoidable harm to those who would benefit from surgery.

12.
bioRxiv (Bioinfo) 2026-06-11

DeePEn - A Depth sensitive benchmark for Protein Engineering

Recent progress in modeling techniques and high-throughput screening has significantly enhanced the accessibility of protein engineering. Nevertheless, further progress gets hindered by the lack of robust benchmarks that capture the practical challenges for real-world protein engineering. Here, we introduced DeePEn, a Depth-sensitive benchmark for Protein Engineering that quantifies a models generalization capabilities when predicting protein fitness at increasing mutational distance from the wildtype or training data. We defined distance as the number of simultaneous point mutations, i.e., single amino acid variants (SAVs), moving from wild-type to mutant (edit distance in computer science jargon). Specifically selecting four deep mutational scanning (DMS) datasets with sufficient multi-mutation data points from ProteinGym, we assessed recent predictive models, including general and biophysics-informed protein Language Models (pLMs), and a non-transformer neural network. Our results highlight how the performance of all models deteriorates with increasing mutational distance and that no single metric sufficiently captures the diverse requirements of protein engineering. To overcome these shortcomings, DeePEn provides a readily available resource for multi-metric benchmarking that focuses on the prediction of distant variants.

13.
arXiv (CS.CL) 2026-06-11

Agreement in Representation Space for Open-Ended Self-Consistency

Self-consistency improves LLM reasoning by sampling multiple outputs and selecting the most consistent answer, but existing formulations largely rely on exact matching and therefore remain limited to tasks with categorical outputs. In this work, we study self-consistency in open-ended generation tasks such as code synthesis and text summarization. We hypothesize that consistency can be understood as a geometric property of the generation space, where semantically compatible generations concentrate in similar regions of representation space. To study this hypothesis, we introduce Embedding-Based Agreement (EBA), a simple training-free operationalization that estimates agreement by clustering sampled generations in embedding space. Through experiments on mathematical reasoning, code generation, and summarization, we show that agreement in representation space provides a robust and scalable signal of self-consistency for open-ended tasks. In particular, EBA consistently outperforms random selection and exhibits more stable scaling behavior than recent selection approaches based on LLM evaluation or uncertainty estimation. We further show that these agreement signals remain stable across model families and embedding spaces, even with native hidden representations. Finally, our analysis shows that the geometric location occupied by sampled generations is strongly correlated with generation quality: generations concentrated near central regions of representation space tend to correspond to more reliable outputs, whereas peripheral generations are substantially less accurate. Overall, our findings support viewing self-consistency as a property of the geometric organization of sampled generations rather than exact symbolic overlap.

14.
arXiv (quant-ph) 2026-06-16

Towards Quantum Limited Spatial Resolution of NV-Diamond Magnetometry

arXiv:2508.13438v2 Announce Type: replace Abstract: Optically addressable ensembles of solid-state defects, such as nitrogen vacancy (NV) centers, are a leading modality for imaging-based magnetometry, thermometry and strain sensing. However, monitoring the fluorescence of individual defects within a sub-diffraction ensemble remains an outstanding challenge that currently limits access to atomic-scale features and dynamics. For compact clusters of NVs, we formulate imaging-based atomic sensing as a low-dimensional multiparameter estimation task in which one seeks to localize each defect and quantify the field strength in its immediate vicinity. In this work, we employ optical spatial mode demultiplexing (SPADE) to enhance localization and brightness estimation accuracy at sub-diffraction scales. Specifically, we develop a two-stage sensing protocol that augments direct imaging by projecting the incoming optical field onto point spread function (PSF)-adapted, i.e., PAD spatial modes and Yuen-Kennedy-Lax (YKL) spatial modes enabling efficient extraction of emitter positions and brightnesses. The YKL-SPADE measurement employed for brightness estimation is shown to be quantum-optimal in the case of two emitters and establishes a new connection between quantum detection and estimation theories. We numerically evaluate the statistical performance of our protocol for sub-diffraction optically detected magnetic resonance (ODMR) and Rabi sensing experiments. Compared to conventional focal plane intensity measurements, our protocol improves emitter localization accuracy by 6$\times$ and brightness estimation accuracy by 2$\times$ for tightly confined ensembles, residing well below the diffraction limit.

15.
arXiv (CS.LG) 2026-06-16

Assessing Predictive Models for Fairness Based on Movement Patterns

arXiv:2605.23234v3 Announce Type: replace Abstract: Assessing the spatial fairness of predictive models involves establishing whether they are statistically penalizing (favoring) individuals associated with certain geographical locations. Literature on this topic makes the fundamental assumption that each individual is assigned to a single geographical location (e.g., place of residence). However, fairness with respect to the set of locations where one has been, i.e., their movement patterns over different regions, also matters when fairness is considered. Consequently, we argue that it is necessary to generalize the notion of spatial fairness to also include movement patterns, leading to the novel problem of assessing predictive models for fairness relative to the movements of individuals. To deal with this problem, we propose an approach that first associates the movements of individuals to certain geographic regions, considering multiple spatial partitions with different resolutions and alignments, and then employs a suitable spatial scan statistic to assess whether a predictive model is fair based on movement patterns. In the experimental evaluation, we study the performance of our approach over thousands of synthetic unfair datasets, showing that it is effective at detecting this new type of unfairness and at retrieving the set of objects treated unfairly, while localization performance exhibits a consistent multi-resolution trade-off.

16.
arXiv (CS.CL) 2026-06-12

LLMs Can Better Capture Human Judgments–With the Right Prompts

Are large language models (LLMs) bad at capturing human judgment? Two commonly stated limitations are that LLMs fail to capture full distributions of responses, and that their judgments are unstable across wording variations. We demonstrate simple prompting strategies that mitigate these limitations. Across two datasets–a U.S.-representative set of 144 moral scenarios and 38 moral beliefs from the International Social Survey Programme's Family and Changing Gender Roles module covering 32 countries–we show how simple elicitation techniques help improve AI-human alignment. First, prompting models to report standard deviations and response proportions recovers the full range of human responses better than common strategies. Second, ensuring scenarios are clear to human participants–as reflected in human confusion ratings–boosts model alignment, and LLMs can track human confusion ratings. At the same time, we find that LLMs' estimates of their own error are poorly calibrated, though they can predict human variability relatively well. These results suggest that asking better questions to LLMs can yield better answers.

17.
arXiv (CS.AI) 2026-06-15

Hyperdimensional computing for structured querying on tabular data embeddings

arXiv:2606.13871v1 Announce Type: new Abstract: Tabular data embeddings have become a cornerstone of data profiling and data integration pipelines, enabling tasks such as entity annotation and resolution; schema matching; column type detection; and table search, among others. Existing approaches embed rows, columns, or entire tables into a vector space and rely on nearest-neighbor search to retrieve candidate matches. A fundamental limitation of current embedding methods is the lack of interpretable similarity scores: the concrete similarity value between a query and its nearest neighbour carries no intrinsic meaning, making it impossible to determine whether that neighbour is a true match or simply the least-dissimilar item in a corpus that contains no valid answer. This inability to set principled thresholds for retrieval undermines practical deployment, particularly for zero-match detection. We investigate the use of HyperDimensional Computing (HDC), specifically the Holographic Reduced Representations (HRR) model, as a framework for tabular row embeddings when the retrieval task corresponds to answering structured select-project queries in vector space. Exploiting the algebraic properties of HDC operations, we derive closed-form expected similarity values for both equality and non-equality retrieval predicates, which converge to interpretable values as dimensionality increases, and use these to identify suitable retrieval thresholds. We evaluate HDC against EmbDI, a graph-based baseline, on two real-world datasets across varying table sizes and predicate lengths. Our results show that HDC matches or outperforms EmbDI for row retrieval across all configurations, handles non-equality predicates more robustly, and achieves perfect attribute projection accuracy at sufficient dimensionality – while uniquely enabling reliable identification of zero-match predicates through its principled thresholds.

18.
arXiv (CS.LG) 2026-06-11

Least-Action-Guided Diffusion for Physical Extrapolation

arXiv:2606.11277v1 Announce Type: new Abstract: Reliable extrapolation remains a central challenge for generative models in computational physics, because models trained over finite ranges of time, parameters, or geometries may produce physically inconsistent predictions outside the training distribution. We introduce a least-action-principle-guided diffusion, LAPG, a framework that promotes physical consistency during inference rather than relying solely on constraints imposed during training. The method combines a conditional score-based diffusion model with an action-derived physical guidance score. In the first stage, the learned score model generates an in-distribution proposal; in the second, an action-based variational prior refines this proposal toward the target out-of-distribution condition. This formulation turns the principle of least action into a differentiable inference-time correction mechanism and provides an alternative to pointwise residual penalties that often require empirical loss balancing. We evaluate LAPG on representative ordinary- and partial-differential-equation systems, including free fall, conservative and dissipative spring-mass dynamics, interacting point vortices, and potential flow over parameterized airfoils. In temporal, parameter, and geometric extrapolation tests, LAPG reduces phase drift, preserves dissipative decay, captures vortex motion, and improves the lift response of airfoil flows compared with training-time physics-informed baselines.

19.
arXiv (CS.LG) 2026-06-18

QUIVER: Cost-Aware Adaptive Preference Querying in Surrogate-Assisted Evolutionary Multi-Objective Optimization

arXiv:2605.04267v2 Announce Type: replace Abstract: Interactive multi-objective optimization systems face a budget allocation dilemma: one can spend resources on expensive objective evaluations or on eliciting decision-maker preferences that identify the relevant region of the Pareto set. Moreover, preference elicitation itself spans modalities with different information content and cognitive burden, ranging from cheap, noisy pairwise preference statements (PS) to richer but costlier indifference adjustments (IA). We study cost-aware optimization under an unknown scalarization and introduce QUIVER (Query-Informed Value Estimation for Regret), a surrogate-assisted evolutionary multi-objective optimizer that adaptively chooses between objective evaluations and heterogeneous preference queries. At each step, QUIVER selects the next action by maximizing the expected decision-quality improvement per unit total cost. Across DTLZ and WFG benchmarks under synthetic decision-maker models, QUIVER achieves the lowest final utility regret on challenging WFG problems (utility regret of 2.14 on WFG4, 2.82 on WFG9: a 25% improvement over baselines), outperforming all single-modality baselines. We analyze how the optimal mix of PS and IA adapts to problem difficulty: on easy problems (DTLZ2), QUIVER selects 80\% PS queries; on hard problems (WFG9), it shifts to 35% IA queries. This adaptive modality selection demonstrates cost-aware preference learning in action.

20.
arXiv (CS.LG) 2026-06-18

On the Stability of Nonlinear Dynamics in GD and SGD: Beyond Quadratic Potentials

arXiv:2602.14789v2 Announce Type: replace Abstract: The dynamical stability of the iterates during training plays a key role in determining the minima obtained by optimization algorithms. For example, stable solutions of gradient descent (GD) correspond to flat minima, which have been associated with favorable features. While prior work often relies on linearization to determine stability, it remains unclear whether linearized dynamics faithfully capture the full nonlinear behavior. Recent work has shown that GD may stably oscillate near a linearly unstable minimum and still converge once the step size decays, indicating that linear analysis can be misleading. In this work, we explicitly study the effect of nonlinear terms. Specifically, we derive an exact criterion for stable oscillations of GD near minima in the multivariate setting. Our condition depends on high-order derivatives, generalizing existing results. Extending the analysis to stochastic gradient descent (SGD), we show that nonlinear dynamics can diverge in expectation even if a single batch is unstable. This implies that stability can be dictated by a single batch that oscillates unstably, rather than an average effect, as linear analysis suggests. Finally, we prove that if all batches are linearly stable, the nonlinear dynamics of SGD are stable in expectation.

21.
arXiv (CS.AI) 2026-06-24

ASymPO: Asymmetric-Scale Policy Optimization for Asynchronous LLM Post-Training Without Behavior Information

arXiv:2606.03070v3 Announce Type: replace-cross Abstract: Asynchronous reinforcement learning can improve language-model post-training throughput by decoupling response generation from policy optimization, but stale responses introduce distribution drift. Standard behavior-corrected methods control this drift with behavior-policy probabilities, importance ratios, or clipping, which requires token-aligned, versioned, and numerically consistent behavior log-probabilities across rollout and learner systems. We ask whether asynchronous group-relative RL can instead be stabilized using only current-policy probabilities. We identify a scale-imbalance failure mode: when stale responses are evaluated under the current policy, positive and negative loss terms can appear at different negative log-probability scales, so zero-sum advantages no longer imply balanced loss contributions. We propose Asymmetric-Scale Policy Optimization (ASymPO), which normalizes each response's token loss by its current average token negative log-probability. ASymPO requires no behavior-policy probabilities, restores response-level zero-sum balance, and preserves a nonzero learning signal. We also introduce Scaled Policy Optimization (SPO), a fixed negative-scaling baseline, and evaluate both current-policy-only objectives in asynchronous mathematical reasoning post-training.

22.
arXiv (CS.AI) 2026-06-12

Examining the Usage of Generative AI Models in Student Learning Activities for Software Programming

arXiv:2511.13271v2 Announce Type: replace-cross Abstract: The rise of Generative AI (GenAI) tools like ChatGPT has created new opportunities and challenges for computing education. Existing research has primarily focused on GenAI's ability to complete educational tasks and its impact on student performance, often overlooking its effects on knowledge gains. In this study, we investigate how GenAI assistance compares to conventional online resources in supporting knowledge gains across different proficiency levels. We conducted a controlled user experiment with 24 undergraduate students of two different levels of programming experience (beginner, intermediate) to examine how students interact with ChatGPT while solving programming tasks. We analyzed task performance, conceptual understanding, and interaction behaviors. Our findings reveal that generating complete solutions with GenAI significantly improves task performance, especially for beginners, but does not consistently result in knowledge gains. Importantly, usage strategies differ by experience: beginners tend to rely heavily on GenAI toward task completion often without knowledge gain in the process, while intermediates adopt more selective approaches. We find that both over-reliance and minimal use result in weaker knowledge gains overall. Based on our results, we call on students and educators to adopt GenAI as a learning rather than a problem solving tool. Our study highlights the urgent need for guidance when integrating GenAI into programming education to foster deeper understanding.

23.
arXiv (CS.AI) 2026-06-15

Refusal Beyond a Single Direction: A Preliminary Comparison of Diff-in-Means and INLP

arXiv:2606.13720v1 Announce Type: new Abstract: Arditi et al. (2024) has shown that refusal in safety fine-tuned chat models is mediated by a single linear direction in the residual stream, recoverable by a difference-in-means (DiM) of harmful and harmless activations. We compare DiM-based interventions (activation addition and directional ablation) with two interventions derived from Iterative Nullspace Projection (INLP) – nullspace projection and counterfactual flipping – on five open-weight chat models, asking whether INLP can match DiM at steering refusal and whether its richer parameterisation yields more tweakable interventions. INLP counterfactual flipping is competitive with DiM directional ablation on refusal suppression, while nullspace projection is consistently weaker. Restricting INLP to the leading directions of the extracted subspace preserves most of the suppression effect at near-baseline perplexity, giving a tunable capability. Geometrically, the two INLP interventions land in qualitatively different regions of activation space: nullspace projection collapses transformed activations between the harmful and harmless clusters, while counterfactual flipping moves them into the opposite cluster, suggesting that the model encodes the absence of a concept differently from its opposite – an intriguing distinction that warrants further investigation in future work.

24.
arXiv (CS.CL) 2026-06-11

Unifying Learning Dynamics and Generalization in Transformers Scaling Law

Authors:

The scaling law, a cornerstone of Large Language Model (LLM) development, predicts improvements in model performance with increasing computational resources. Yet, while empirically validated, its theoretical underpinnings remain poorly understood. This work formalizes the learning dynamics of transformer-based language models as an ordinary differential equation (ODE) system, then approximates this process to kernel behaviors. Departing from prior toy-model analyses, we rigorously analyze stochastic gradient descent (SGD) training for multi-layer transformers on sequence-to-sequence data with arbitrary data distribution, closely mirroring real-world conditions. Our analysis characterizes the convergence of generalization error to the irreducible risk as computational resources scale with data, especially during the optimization process. We establish matching upper and lower bounds on the excess risk, characterized by a distinct phase transition. In the initial optimization phase, the excess risk decays exponentially relative to the computational cost ${\sf C}$. However, once a specific resource allocation threshold is crossed, the system enters a statistical phase, where the generalization error follows a power-law decay of $\Theta(\mathsf{C}^{-1/7})$. These rates are certified by complementary lower bounds – statistical, via an information-theoretic two-point reduction, and optimization-side, via a first-order oracle argument – rendering the two-stage law tight up to constants, logarithmic factors, and a condition-number gap. Beyond this unified framework, our theory derives isolated scaling laws for model size, training time, and dataset size, elucidating how each variable independently governs the bounds of generalization.

25.
arXiv (CS.CL) 2026-06-24

Self-Recognition Finetuning can Prevent and Reverse Emergent Misalignment

Emergent misalignment (EM) has been linked to the activation of misaligned persona vectors and evil character traits, suggesting that EM operates through disruption of the model's aligned character rather than direct learning of harmful content. Motivated by this connection, we study self-generated text recognition (SGTR) finetuning as a character-targeted intervention that is distinct from existing in-training defenses. We conduct two-stage finetuning experiments across three models (GPT-4.1, Qwen2.5-32B-Instruct, Seed-OSS-36B-Instruct) and multiple EM datasets to compare SGTR finetuning against benign finetuning baselines (correct domain-specific data, general knowledge, and word counting) to find it an effective defense in both reversal and prevention settings. We find that all interventions produce comparable EM reversal, but only when restoring capabilities that EM had degraded. For prevention, only SGTR finetuning consistently reduces misalignment without exacerbating any individual metric, suggesting that character fortification specifically drives prevention. We provide further evidence for EM's relation to the LLM's default character by showing that EM finetuning induces diversity into the LLM's identity self-reports, artificially corrupting self-recognition exacerbates misalignment caused by EM finetuning, and that removing the model's identity-bearing system prompt substantially reduces the effect of EM finetuning. Together, these findings reframe EM not as the adoption of a coherent misaligned persona but as the destabilization of aligned character.