Academic Intelligence · Curated Daily

探索全球前沿学术脉络

AcademicHub 汇聚顶级期刊与预印本平台的实时文献。定制您的专属科研雷达,利用大语言模型自动生成交叉领域文献分析简报。

01.
arXiv (CS.AI) 2026-06-16

Scaling Adaptive Depth with Norm-Agnostic Residual Networks

arXiv:2606.16112v1 Announce Type: cross Abstract: Residual architectures are ubiquitous in deep learning, but they suffer from a subtle structural limitation: the norm of the residual stream can grow rapidly with depth. As a result, updates from later layers become small relative to the accumulated residual state. This reduces their impact on the representation and limits the benefits of scaling models in depth. To address this, we introduce NAG, a norm-agnostic residual architecture that separates magnitude from directional information in the residual stream, preserving meaningful layer contributions throughout depth and preventing later updates from being systematically suppressed by residual-norm growth. Importantly, NAG introduces only a negligible number of additional parameters and relies on simple operations that are easily kernel-fusible, preserving training efficiency in practice. We show that this architecture outperforms baseline Transformers, with gains that increase substantially as depth grows, enabling effective training of much deeper models. The norm-agnostic formulation also leads to an interpretable Mixture-of-Depths (MoD) mechanism that adaptively skips both attention and MLP layers. Beyond serving as a post-training accuracy-compute tradeoff, this mechanism can be used as a pretraining-time scaling strategy: under iso-FLOP training, compute saved by reducing per-token forward-pass cost can be reinvested into training on more tokens while keeping the total parameter count and KV-cache budget fixed. In our experiments, moderate Mixture-of-Depths rates of approximately 20%-25% match full-depth baseline performance under equal training compute while substantially reducing the number of executed layer parameters and forward-pass FLOPs. These results identify sparsity in depth as a new scaling axis for fixed-compute training, enabling very deep yet FLOP-efficient models.

02.
arXiv (CS.AI) 2026-06-12

The Query Channel: Information-Theoretic Limits of Masking-Based Explanations

arXiv:2604.16689v2 Announce Type: replace Abstract: Masking-based post-hoc explanation methods, such as KernelSHAP and LIME, estimate local feature importance by querying a black-box model under randomized perturbations. This paper formulates this procedure as communication over a query channel, where the latent explanation acts as a message and each masked evaluation is a channel use. Within this framework, the complexity of the explanation is captured by the entropy of the hypothesis class, while the query interface supplies information at a rate determined by an identification capacity per query. We derive a strong converse showing that, if the explanation rate exceeds this capacity, the probability of exact recovery necessarily converges to one in error for any sequence of explainers and decoders. We also prove an achievability result establishing that a sparse maximum-likelihood decoder attains reliable recovery when the rate lies below capacity. A Monte Carlo estimator of mutual information yields a non-asymptotic query benchmark that we use to compare optimal decoding with Lasso- and OLS-based procedures that mirror LIME and KernelSHAP. Experiments reveal a range of query budgets where information theory permits reliable explanations but standard convex surrogates still fail. Finally, we interpret super-pixel resolution and tokenization for neural language models as a source-coding choice that sets the entropy of the explanation and show how Gaussian noise and nonlinear curvature degrade the query channel, induce waterfall and error-floor behavior, and render high-resolution explanations unattainable.

03.
arXiv (CS.CV) 2026-06-19

Shape of Thought: Progressive Object Assembly via Visual Chain-of-Thought

Multimodal models for text-to-image generation have achieved strong visual fidelity, yet they remain brittle under compositional structural constraints, notably generative numeracy, attribute binding, and part-level relations. To address these challenges, we propose Shape-of-Thought (SoT), a visual CoT framework for process-supervised progressive shape assembly in the rendered 2D domain, without external engines at inference time. SoT trains a unified multimodal autoregressive model to generate interleaved textual plans and rendered intermediate states, helping the model capture shape-assembly logic without producing explicit geometric representations. Unlike text-only CoT, each decision is grounded in a rendered state, making counts, attachments, topology, and intermediate part-addition errors inspectable across the trajectory. To support this paradigm, we introduce SoT-26K, a large-scale dataset of grounded assembly traces derived from part-based CAD hierarchies, and T2S-CompBench, a benchmark for evaluating structural integrity and trace faithfulness. Fine-tuning on SoT-26K achieves 88.4% on component numeracy and 84.8% on structural topology, outperforming direct generation by +24.2 points on component numeracy and +19.3 points on structural topology. SoT establishes a transparent testbed for rendered-domain structure-aware generation. The code is available at https://github.com/yuhuo03/Shape-of-Thought.

04.
arXiv (CS.CL) 2026-06-18

ActMem: Bridging the Gap Between Memory Retrieval and Reasoning in LLM Agents

Memory management is essential for LLM agents in long-term interactions. Current memory frameworks typically treat agents as passive ``recorders'' and retrieve information without understanding its deeper implications. They may fail in scenarios requiring reasoning and complex decision-making. To bridge this critical gap, we propose a novel actionable memory framework called ActMem that integrates memory retrieval with active causal reasoning. ActMem transforms unstructured dialogue history into a structured causal and semantic graph. By leveraging counterfactual reasoning and commonsense completion, it enables agents to deduce implicit constraints and resolve potential conflicts between past states and current intentions. Furthermore, we introduce a comprehensive dataset ActMemEval to evaluate agent reasoning capabilities in logic-driven scenarios, moving beyond the fact-retrieval focus of existing memory benchmarks. Experiments demonstrate that ActMem significantly outperforms baselines in handling complex, memory-dependent tasks, paving the way for more consistent and reliable intelligent assistants.

05.
arXiv (CS.AI) 2026-06-12

Exploring How Agent Voice Accents Shape Human-AI Collaboration in K-12 Group Learning

arXiv:2606.12805v1 Announce Type: cross Abstract: Collaboration is widely recognized as a cornerstone of 21st-century education, yet teachers still encounter persistent challenges in fostering productive peer interaction. LLM conversational peer agents introduce new possibilities for mediating in-person group work, raising questions about how persona design, particularly their voice characteristics, shapes learners' perceptions, trust, and interactional dynamics. While prior work has examined agent accent effects in one-to-one settings, little is known about how these effects manifest in groups. We conducted a between-subjects mixed-methods study with 33 teachers examining how a GenAI voice agent with different accents (British, Indian, and African American) influenced collaboration and agent perception. Across surveys, group interaction analyses, and artifacts, we find that accent shaped participants' mental models and the roles the agent assumed in group interaction. The British-accented agent was largely treated as a tool and engaged in detached, utility-based ways, whereas Indian- and African American-accented agents were more readily anthropomorphized and integrated as peers. These role expectations influenced trust, engagement, and reliance over time. This work advances understanding of how GenAI's sociolinguistic design features shape group dynamics in CSCL, with implications for designing culturally inclusive AI partners in group learning.

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

MegaFold: Efficient Training of Next-Generation 3D Attention Protein Models on Cross-Platform GPUs

arXiv:2506.20686v2 Announce Type: replace-cross Abstract: Recent advances in biomolecular modeling have been catalyzed by models such as AlphaFold3 (AF3), which introduce science-informed changes to the transformer architecture. Unlike transformers, a defining characteristic of AF3-style models is their 3D attention over 2D pairwise representations which produces tensors whose computation and memory costs scale cubically with sequence length. As a result, despite moderate parameter counts, AF3-style models are far more expensive to train than size-equivalent transformers, and are severely constrained by GPU memory capacity. Our characterization shows 3D attention fundamentally changes the training workload, causing massive 3D attention maps, complex inter-operator dependencies, kernel fragmentation, and heavy host-side data pipelines which differ substantially from LLM training, leading to poor utilization on modern GPU systems. Moreover, existing GPU optimizations do not adequately address these challenges due to complex cross-layer inter-operator dependencies introduced by 3D attention. Motivated by these challenges, we introduce MegaFold, a novel cross-platform system for efficient training of next-generation 3D-attention protein models. MegaFold combines a memory-efficient 3D-attention kernel, a communication-efficient sharding strategy for quadratic representations, fused operator implementations for critical execution paths, and a determinism-aware host-device pipeline that eliminates preprocessing stalls. Evaluation on both NVIDIA H200 and AMD MI250 GPUs shows that MegaFold enables training with up to 3.36$\times$ longer sequence lengths on 32 GPUs while reducing end-to-end execution time by up to 1.73$\times$ (NVIDIA) and 1.62$\times$ (AMD).

07.
arXiv (CS.AI) 2026-06-16

MBABench: Evaluating LLM Agents on End-to-End Spreadsheet Tasks in Finance

arXiv:2605.22664v3 Announce Type: replace Abstract: LLM agents are increasingly expected to carry out end-to-end workflows, producing complete artifacts from high-level user instructions. To meet enterprise needs, frontier AI labs have developed agents that can construct entire spreadsheets from scratch. This is especially relevant in finance, where core workflows such as financial modeling, forecasting, and scenario analysis are commonly conducted through spreadsheets. Yet, existing spreadsheet benchmarks do not measure this advanced capability, focusing instead on question-answering or single-formula edits. To address this gap, we provide one of the first evaluations of agents on end-to-end spreadsheet tasks, focusing on economically critical financial workflows such as modeling and scenario analysis. Since deliverables therein are routinely reviewed and revised by multiple stakeholders, judging their quality necessarily involves high-level criteria such as readability or ease of modification. To reflect the multidimensional nature of solution quality, we develop an evaluation taxonomy comprising three dimensions: Accuracy, Formula, and Format, each comprising fine-grained criteria that reflect professional standards. The Claude family leads the benchmark and produces the most professional-looking outputs in our qualitative review, but even the strongest agents frequently fall short of professional finance standards and degrade sharply as the difficulty increases beyond a few chained calculations. This suggests that current agents are not yet able to reliably produce professional-quality spreadsheets at the level of complexity real-world workflows demand.

08.
arXiv (CS.CL) 2026-06-19

Generative Engine Optimization at Scale: Measuring Brand Visibility Across AI Search Engines

People increasingly get answers straight from AI search engines like ChatGPT, Claude, Perplexity, and Gemini rather than scrolling search results. Brands that once focused on search engine optimization (SEO) must now optimize for how these engines represent, cite, and recommend them – a shift variously called Generative Engine Optimization (GEO), Answer Engine Optimization (AEO), and AI Search Visibility. We treat AEO and AI Visibility as part of GEO, and study how to measure brand visibility across AI engines: what they value when they cite a brand, which sources they rely on, and what content large language models surface. The hard case is everyone outside the already-authoritative top brands – SMEs, D2C brands, creators, and early-stage startups. We analyze 100K+ prompt responses across 100+ brands tracked on Ranqo between March and May 2026. First visibility runs form a clear three-tier brand-stature ladder: global household names (e.g., Stripe, Nike) appear in 73% of relevant AI answers on their first run; established mid-market and regional brands (e.g., Olipop, Klaviyo) in 44%; niche and small brands in just 11% – about 30 percentage points per step. When engines cite sources, about 78% go to corporate websites; among non-corporate sources YouTube leads, ahead of Reddit, editorial media, and Wikipedia. The highest-leverage page is the ranked "best-of" listicle, the most-cited content format at about 21% of all citations. Sentiment is the unstable signal: whether a brand is framed positively or negatively flips about 6.7 times more often than whether it is mentioned at all. These findings provide a first large-scale baseline for measuring GEO: AI brand visibility can be measured, differs by platform, and varies strongly by brand maturity. We close by proposing seven v1.1 protocols to test whether specific recommendations can causally improve AI visibility.

09.
arXiv (CS.LG) 2026-06-12

From Uncertain Judgments to Calibrated Rankings: Conformal Elo Estimation for LLM Evaluation

arXiv:2606.13221v1 Announce Type: new Abstract: Evaluating new large language models typically requires costly human annotation campaigns at scale. LLM-as-a-judge offers a cheaper alternative, but judge scores carry systematic errors - such as position bias, self-preference, or intransitivity - that can strongly miscalibrate the resulting rankings. We quantify the resulting judge-human disagreement at two complementary levels. At the local level, we estimate per-battle uncertainty from the judge's own score differences by propagating calibrated win probabilities rather than hard labels into the Bradley-Terry procedure. This alone provides a drastic improvement to Elo estimation accuracy, bringing LLM-derived ratings within 17.9 Elo MAE of human-derived ones when averaged over 55 held-out models on LMArena. At the global level, we apply split conformal prediction to the residual gap between LLM-derived and human-derived Elo ratings across held-out models, producing prediction intervals with distribution-free marginal coverage guarantees that account for irreducible LLM-human disagreement. Together, these two layers yield a low-cost evaluation tool that provides developers with calibrated Elo estimates and honest uncertainty bounds, without access to large-scale human annotations.To facilitate reproducibility, we release our code at https://github.com/kargibora/SoftElo .

10.
arXiv (CS.AI) 2026-06-15

UltraSketchLLM: Sub-1-Bit LLM Compression via Sketch and Hardware-Friendly Operators

arXiv:2506.17255v2 Announce Type: replace-cross Abstract: Large language models (LLMs) require larger GPU memory size these days, necessitating efficient and extreme weight compression methods. Existing compression methods are either theoretically limited by 1 bit per weight or face severe performance degradation and inefficiency. To deploy LLMs in resource-constrained scenarios, we introduce UltraSketchLLM, compressing LLMs with data sketch. It reduces peak GPU memory footprint with a high compression rate down to 0.5 bit per weight. Combined with hardware-friendly implementation, UltraSketchLLM keeps tolerable performance degradation and extremely low latency overhead with 14.9x speedup compared to naive sketch solution.

11.
arXiv (math.PR) 2026-06-19

Establishing an $\Omega(\sqrt{d})$ complexity lower bound for PDMP samplers and how to break it: a sub-$\sqrt{d}$ algorithm for Gaussian-tailed targets

arXiv:2606.19909v1 Announce Type: cross Abstract: Despite the theoretical appeal of their non-reversibility, to date, no Piecewise Deterministic Markov Process (PDMP) samplers have been developed that scale better than $\mathcal{O}(\sqrt{d})$ in computational complexity with respect to the target dimension $d$. We prove that this is a fundamental limitation by establishing an $\Omega(\sqrt{d})$ lower bound on the algorithmic complexity of PDMP samplers in a standard setup. By relaxing the assumption that the target density must remain invariant at all continuous times, we then demonstrate how to bypass this barrier. Specifically, we introduce a novel PDMP sampling scheme and show that it achieves an empirical complexity of $\mathcal{O}(d^\alpha)$, where $\alpha \in [0.2, 0.3]$ for Gaussian-tailed targets. In addition, this PDMP scheme is locally adaptive in both trajectory length and distance between velocity updates.

12.
arXiv (CS.LG) 2026-06-12

Prediction-Powered Causal Inference by Automatic Debiased Machine Learning and Semi-Supervised Riesz Regression

arXiv:2606.12892v1 Announce Type: cross Abstract: This study investigates semiparametric efficient estimation of causal and structural parameters in a semi-supervised setting. In our setting, unlabeled auxiliary regressors are available in addition to labeled observations consisting of outcomes and regressors. Our goal is to construct estimators of causal and structural parameters whose asymptotic variances are smaller than those of estimators constructed using only labeled data. We refer to this framework as prediction-powered causal inference (PPCI). We first derive the efficient influence function and the efficiency bound, which imply that the use of auxiliary regressors can attain a smaller asymptotic variance than the efficiency bound attainable from labeled observations alone. Then, by combining the efficient influence function with the debiased machine learning (DML) framework, we propose methods that we call DML-PPCI. If we construct an estimating-equation estimator, we refer to the method as EE-DML-PPCI; if we construct a targeted-learning estimator, we refer to the method as TMLE-DML-PPCI. The asymptotic variances of both estimators match our derived efficiency bound. In the construction of the estimators, estimation of the efficient influence function plays an important role. In our study, the efficient influence function is also a Neyman orthogonal score, which depends on the Riesz representer and the regression function. For Riesz representer estimation, we develop semi-supervised generalized Riesz regression with convergence rate guarantees.

13.
arXiv (CS.LG) 2026-06-18

Self-attention-based non-linear basis transformations for compact latent space modelling of dynamic optical fibre transmission matrices

arXiv:2406.07775v2 Announce Type: replace Abstract: Multimode optical fibres are hair-thin strands of glass that efficiently transport light. They promise next-generation medical endoscopes that provide unprecedented sub-cellular image resolution deep inside the body. However, confining light to such fibres means that images are inherently scrambled in transit. Conventionally, this scrambling has been compensated by pre-calibrating how a specific fibre scrambles light and solving a stationary linear matrix equation that represents a physical model of the fibre. However, as the technology develops towards real-world deployment, the unscrambling process must account for dynamic changes in the matrix representing the fibre's effect on light, due to factors such as movement and temperature shifts, and non-linearities resulting from the inaccessibility of the fibre tip when inside the body. Such complex, dynamic and nonlinear behaviour is well-suited to approximation by neural networks, but most leading image reconstruction networks rely on convolutional layers, which assume strong correlations between adjacent pixels, a strong inductive bias that is inappropriate for fibre matrices which may be expressed in a range of arbitrary coordinate representations with long-range correlations. We introduce a new concept that uses self-attention layers to dynamically transform the coordinate representations of varying fibre matrices to a basis that admits compact, low-dimensional representations suitable for further processing. We demonstrate the effectiveness of this approach on diverse fibre matrix datasets. We show our models significantly improve the sparsity of fibre bases in their transformed bases with a participation ratio, p, as a measure of sparsity, of between 0.01 and 0.11. Further, we show that these transformed representations admit reconstruction of the original matrices with < 10% reconstruction error, demonstrating the invertibility.

14.
arXiv (CS.CL) 2026-06-16

OBCache: Optimal Brain KV Cache Pruning for Efficient Long-Context LLM Inference

Large language models (LLMs) with extended context windows enable powerful applications but impose significant memory overhead, as caching all key-value (KV) states scales linearly with sequence length and batch size. Existing cache eviction methods address this by exploiting attention sparsity, yet they typically rank tokens heuristically using accumulated attention weights without considering their true impact on attention outputs. We propose Optimal Brain Cache (OBCache), a principled framework that formulates cache eviction as a layer-wise structured pruning problem. Building upon the Optimal Brain Damage (OBD) theory, OBCache quantifies token saliency by measuring the perturbation in attention outputs induced by pruning tokens, with closed-form scores derived for isolated keys, isolated values, and joint key-value pairs. Our scores account not only for attention weights but also for information from value states and attention outputs, thereby enhancing existing eviction strategies with output-aware signals. Experiments on LLaMA and Qwen models demonstrate that replacing the heuristic scores in existing works, which estimate token saliency across different query positions, with OBCache's output-aware scores consistently improves long-context accuracy. Code is available at https://github.com/DreamSoul-AI/OBCache.

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

KGEdit: Ambiguity-Aware Knowledge Graphs for Training-Free Precise Video Generation and Editing

In recent years, training-free video generation has progressed remarkably. However, when handling complex textual instructions, existing methods still suffer from semantic ambiguity, incorrect concept binding, and cross-frame inconsistency. To address these issues, we propose KGEdit, a structured semantic control framework for text-to-video (T2V) diffusion models. Specifically, we first construct an ambiguity-aware knowledge graph (AAKG) to disentangle and disambiguate the input prompt, converting it into four types of structured semantics: identity, relation, attribute, and negative constraints. We then design a structured semantic injection module (SSIM) to inject these semantic signals into key layers of the diffusion Transformer, enabling fine-grained semantic control. In addition, we introduce a temporal-aware semantic control (TASC) module that dynamically schedules semantic objectives according to the stage-wise characteristics of the denoising process, further improving semantic alignment and temporal consistency. Experiments show that KGEdit outperforms existing methods in editing precision and temporal stability, while offering higher efficiency and controllability in text-driven interaction scenarios.

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

CD-RCM: Generalizable Continuous-Depth Novel View Synthesis for Reflectance Confocal Microscopy

Reflectance confocal microscopy (RCM) provides noninvasive, cellular-resolution "optical biopsies" of human skin in vivo by acquiring en-face images at successive depths, forming a sparse z-stack. Due to optical limitations, these stacks are anisotropic 3D volumes with lateral resolution (0.5 $\mu$m) $\sim$6 times higher compared to axial resolution, which is defined by the optical sectioning (3 $\mu$m), limiting the interpretation of tissue. Our goal is to provide continuous-depth visualization by interpolating intermediate sections and making the 3D volume isotropic. Such a representation permits arbitrary-direction sectioning, including histopathology-like cross-sectional examination, without requiring per-patient optimization. To that end, we introduce the first RCM-specific novel-view synthesis (NVS) approach, CD-RCM, a feedforward model that predicts realistic, unseen depths from sparsely sampled RCM stacks. Classical neural rendering methods focus on reconstruction from surface-level multi-view observations. In contrast to surface-level camera views, RCM can acquire optically sectioned en-face images of tissue beyond the surface up to 200 $\mu$m. However, during visualization of the RCM stacks, observations of the shallower sections (towards the surface) obscure the deeper ones. This unique axial imaging geometry and layer-dependent anatomical organization motivated our development of a tailored architectural and training framework that explicitly accounts for RCM's depth-resolved, occlusive imaging physics. Experiments demonstrate that CD-RCM achieves high-fidelity novel-view synthesis with sub-second inference time.

17.
arXiv (CS.AI) 2026-06-12

ReCal: Reward Calibration for RL-based LLM Routing

arXiv:2606.12479v1 Announce Type: cross Abstract: Large language model (LLM) routing has emerged as an effective paradigm for leveraging the complementary strengths of multiple LLMs through dynamic model and reasoning-strategy selection. Recent reinforcement learning (RL)-based routing methods further improve routing quality by optimizing routing policies from interaction feedback. However, they still struggle to provide informative and comparable learning signals under heterogeneous tasks with varying difficulty. In practice, multiple objectives (e.g., correctness, format behavior) are aggregated into a single scalar reward, leading to ambiguous credit assignment and conflicting optimization signals. Moreover, reward signals exhibit significant variability across instances, where some instances produce higher or more variable rewards, introducing optimization bias that favors trivial samples over informative ones. To address these issues, we propose ReCal, a \underline{Re}ward \underline{Cal}ibration framework for RL-based LLM routing. We first introduce a hierarchical reward decomposition mechanism with component-wise advantage estimation. We further propose a distribution-aware optimization strategy that calibrates optimization variability through variance-aware reweighting and per-dataset normalization. Experiments on seven datasets demonstrate that ReCal consistently improves routing performance, and training stability over baselines. Code is available at https://anonymous.4open.science/r/ReCal.

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

SSD: Spatially Speculative Decoding Accelerates Autoregressive Image Generation

Autoregressive models excel in visual generation by treating images as 1D sequences of discrete tokens, mirroring language modeling. However, this flattening discards the intrinsic 2D spatial locality of visual signals, creating severe computational bottlenecks during inference. We introduce Spatially Speculative Decoding (SSD), a framework that aligns the predictive objective with the natural geometry of images. Rather than predicting only the immediate next token in a 1D sequence, our model simultaneously predicts the adjacent horizontal token and the token directly below it. By capitalizing on this 2D spatial correlation, spatially speculative decoding overcomes the memory wall in visual inference. Our approach accelerates autoregressive image generation by up to 13.3x while maintaining high fidelity on DPG-Bench and GenEval. Our results suggest that respecting the underlying geometry of vision unlocks massive computational efficiencies, paving the way for real-time, high-resolution autoregressive generative models.

19.
arXiv (CS.CV) 2026-06-19

DiffMath: Symbol- and Graph-Aware Latent Diffusion Transformer for Handwritten Mathematical Expression Generation

Handwritten Mathematical Expression Generation (HMEG) is challenging due to the complex two-dimensional layouts and long-range structural dependencies of mathematical expressions. Existing methods typically rely on explicit spatial supervision, such as symbol-level bounding boxes, which incurs high annotation costs and limits scalability. In this work, we propose DiffMath, a symbol- and graph-aware latent diffusion framework that leverages the hierarchical structure inherent in LaTeX as a structural prior, eliminating the need for positional supervision. First, we design a Relational Abstract Syntax Tree (RelAST), a generation-oriented representation that distills MathML trees into compact triplet sequences [S, R, D], where each token directly encodes a symbol identity, spatial relation, or nesting depth. Second, we introduce MathVAE, which learns structure-preserving latent representations through symbol-aware and relation-aware perceptual regularization, ensuring that the latent space captures both character semantics and spatial topology. Third, MathDiT performs conditional denoising in this structured latent space, further guided by a global symbol-count prior via Adaptive Layer Normalization (AdaLN) to improve structural coherence. Experiments show that DiffMath produces structurally consistent handwritten expressions, achieves superior performance over existing methods, and improves the accuracy of downstream OCR models through synthetic data augmentation.

20.
arXiv (CS.LG) 2026-06-15

IntSeqBERT: Learning Arithmetic Structure in OEIS via Modulo-Spectrum Embeddings

arXiv:2603.05556v2 Announce Type: replace Abstract: Integer sequences in the OEIS span values from single-digit constants to astronomical factorials and exponentials, making prediction challenging for standard tokenised models that cannot handle out-of-vocabulary values or exploit periodic arithmetic structure. We present IntSeqBERT, a dual-stream Transformer encoder for masked integer-sequence modelling on OEIS. Each sequence element is encoded along two complementary axes: a continuous log-scale magnitude embedding and sin/cos modulo embeddings for 100 residues (moduli $2$–$101$), fused via FiLM. Three prediction heads (magnitude regression, sign classification, and modulo prediction for 100 moduli) are trained jointly on 274,705 OEIS sequences. At the Large scale (91.5M parameters), IntSeqBERT achieves 95.85% magnitude accuracy and 50.38% Mean Modulo Accuracy (MMA) on the test set, outperforming a standard tokenised Transformer baseline by $+8.9$ pt and $+4.5$ pt, respectively. An ablation removing the modulo stream confirms it accounts for $+15.2$ pt of the MMA gain and contributes an additional $+6.2$ pt to magnitude accuracy. A probabilistic Chinese Remainder Theorem (CRT)-based Solver converts the model's predictions into concrete integers, yielding a 7.4-fold improvement in next-term prediction over the tokenised-Transformer baseline (Top-1: 19.09% vs. 2.59%). Modulo spectrum analysis reveals a strong negative correlation between Normalised Information Gain (NIG) and Euler's totient ratio $\varphi(m)/m$ ($r = -0.851$, $p < 10^{-28}$), providing empirical evidence that composite moduli capture OEIS arithmetic structure more efficiently via CRT aggregation.

21.
arXiv (CS.CL) 2026-06-19

GEMS: Geometric Constraints Enable Multi-Semantic Superposition in LLMs

作者:

Activation steering controls model behavior by modifying intermediate hidden states at inference time without retraining. Existing methods handle only single-direction injection; when multiple semantic directions are superposed without constraints, the model collapses. We show that this collapse decomposes into two independently acting sources: distributional deviation, where additive perturbations accumulate in norm across layers and drive activations outside the training distribution, and directional interference, where non-orthogonal semantic vectors mutually dampen when superposed. These two sources define the design constraints that any training-free multi-directional intervention must address. As one instantiation of these principles, we propose GEMS, a training-free method that maps each source to a corresponding geometric constraint: norm-preserving weighted superposition and targeted attention-pathway injection for distributional deviation, and real-time orthogonalization for directional interference. On GSM8K, injecting three concurrent non-mathematical directions preserves accuracy at 98% (baseline 92%), while unconstrained addition collapses to 4%; on Wikitext-2, the same injection incurs only 2.2% PPL increase. Component ablation isolates the causal role of each constraint, and layer-level probes confirm that orthogonalized signals survive the FFN pathway and reach the output distribution with semantic specificity. Qualitative steering effects transfer across architectures from 3B to 31B.

22.
arXiv (CS.CL) 2026-06-16

daVinci-kernel: Co-Evolving Skill Selection, Summarization, and Utilization via RL for GPU Kernel Optimization

GPU kernel optimization represents a paradigm where functional correctness is assumed and execution efficiency is the objective. We present daVinci-kernel, a reinforcement learning framework that couples skill discovery with skill exploitation through a dynamically evolving skill library. daVinci-kernel jointly trains three agents sharing one LLM backbone: a Skill Selection Agent that retrieves relevant techniques via BM25 and LLM reranking, a Policy Agent that generates multi-turn CUDA/Triton kernels conditioned on selected skills, and a Skill Summary Agent that distills successful rollouts into reusable skills. Candidate skills are added only after execution-based verification confirms reproducible speedups. All three agents share a single LLM backbone, are initialized via a structured SFT cold start on diversity-filtered data, and are then jointly optimized end-to-end with multi-turn REINFORCE and per-agent advantage estimation. On KernelBench, daVinci-kernel-14B achieves 37.2%, 70.6%, and 32.2% on Level 1, Level 2, and Level 3 under the Fast$_1$ threshold, outperforming the strongest prior RL-trained model, Dr.Kernel-14B.

23.
arXiv (CS.AI) 2026-06-16

Integrating Reasoning and Generalization in Text-to-SQL via Self-Enhanced Fine-Tuning

arXiv:2606.15598v1 Announce Type: new Abstract: Text-to-SQL aims to translate natural language questions into executable SQL queries over structured databases, enabling non-expert users to access data intuitively. While recent advances in large language models (LLMs) have shown promise in this task, existing LLM-based approaches often struggle to strike a balance between strong reasoning capabilities and robust generalization. To address these limitations, we propose CoTE-SQL to enhance the LLM-based text-to-SQL generation with three key innovations: (i) self-enhanced reasoning traces distilled from LLMs without human annotation, (ii) structured chain-of-thought (CoT) prompting with modular decomposition and examples retrieval, and (iii) error-aware revision based on SQL execution feedback. Extensive experiments on the Spider and Bird benchmarks demonstrate that CoTE-SQL achieves new state-of-the-art performance among methods built on open-source LLMs with comparable model sizes on Bird (53.39% EX / 59.02 VES) and strong results on Spider (79.60% EX / 77.19 VES), with especially significant gains on complex queries. Results highlight the effectiveness of combining self-enhancement, structured reasoning, and execution-time feedback within an LLM-based framework for text-to-SQL design.

24.
arXiv (CS.CV) 2026-06-16

Random Erasing vs. Model Inversion: A Promising Defense or a False Hope?

Model Inversion (MI) attacks pose a significant privacy threat by reconstructing private training data from machine learning models. While existing defenses primarily concentrate on model-centric approaches, the impact of data on MI robustness remains largely unexplored. In this work, we explore Random Erasing (RE), a technique traditionally used for improving model generalization under occlusion, and uncover its surprising effectiveness as a defense against MI attacks. Specifically, our novel feature space analysis shows that models trained with RE-images introduce a significant discrepancy between the features of MI-reconstructed images and those of the private data. At the same time, features of private images remain distinct from other classes and well-separated from different classification regions. These effects collectively degrade MI reconstruction quality and attack accuracy while maintaining reasonable natural accuracy. Furthermore, we explore two critical properties of RE including Partial Erasure and Random Location. Partial Erasure prevents the model from observing entire objects during training. We find this has a significant impact on MI, which aims to reconstruct the entire objects. Random Location of erasure plays a crucial role in achieving a strong privacy-utility trade-off. Our findings highlight RE as a simple yet effective defense mechanism that can be easily integrated with existing privacy-preserving techniques. Extensive experiments across 37 setups demonstrate that our method achieves state-of-the-art (SOTA) performance in the privacy-utility trade-off. The results consistently demonstrate the superiority of our defense over existing methods across different MI attacks, network architectures, and attack configurations. For the first time, we achieve a significant degradation in attack accuracy without a decrease in utility for some configurations.

25.
arXiv (CS.AI) 2026-06-19

Bi-Anchor Interpolation Solver for Accelerating Generative Modeling

arXiv:2601.21542v3 Announce Type: replace-cross Abstract: Flow Matching (FM) models have emerged as a leading paradigm for high-fidelity synthesis. However, their reliance on iterative Ordinary Differential Equation (ODE) solving creates a significant latency bottleneck. Existing solutions face a dichotomy: training-free solvers suffer from significant performance degradation at low Neural Function Evaluations (NFEs), while training-based one- or few-steps generation methods incur prohibitive training costs and lack plug-and-play versatility. To bridge this gap, we propose the Bi-Anchor Interpolation Solver (BA-solver). BA-solver retains the versatility of standard training-free solvers while achieving significant acceleration by introducing a lightweight SideNet (1-2% backbone size) alongside the frozen backbone. Specifically, our method is founded on two synergistic components: 1) Bidirectional Temporal Perception, where the SideNet learns to approximate both future and historical velocities without retraining the heavy backbone; and 2) Bi-Anchor Velocity Integration, which utilizes the SideNet with two anchor velocities to efficiently approximate intermediate velocities for batched high-order integration. By utilizing the backbone to establish high-precision ``anchors'' and the SideNet to densify the trajectory, BA-solver enables large interval sizes with minimized error. Empirical results on ImageNet-256^2 demonstrate that BA-solver achieves generation quality comparable to 100+ NFEs Euler solver in just 10 NFEs and maintains high fidelity in as few as 5 NFEs, incurring negligible training costs. Furthermore, BA-solver ensures seamless integration with existing generative pipelines, facilitating downstream tasks such as image editing.