arXiv subjects

Huan Wang

Publications and source records attributed to Huan Wang.

At least 19 recordsLinked to original sources

arXiv2026

Is Oracle Pruning the True Oracle?

Oracle pruning, which selects unimportant weights by minimizing the pruned train loss, has served as the foundation for most neural network pruning methods for over thirty-five years, while few (if any) have thought about how much the foundation really holds. This paper, for the first time, attempts to systematically examine its validity on deep neural networks through empirical correlation analyses and provides meta-framework reflections on the field of neural network pruning. Specifically, this paper focuses on the pruning algorithms with three stages: training, pruning, and retraining. We analyze the correlation in model performance before and after the retraining stage. Extensive experiments (37K models are trained) across a wide spectrum of models (LeNet5, VGG, ResNets, ViT, MLLM) and datasets (MNIST, CIFAR10/CIFAR100, ImageNet-1K, MLLM data) are conducted. For large-scale experiments, we adopt approximate oracle pruning due to the prohibitive cost of exact oracle pruning. The results point to a counterintuitive conclusion: for deep learning models of nontrivial size (already at the scale of ResNet56 on CIFAR-10), pre-retraining performance is negligibly correlated with post-retraining performance. In other words, the weights identified by oracle pruning can scarcely guarantee strong performance following retraining. This further suggests that existing works that derive pruning criteria from oracle pruning may rest on a questionable foundational premise. Further studies suggest that rising task complexity is a primary factor behind the invalidity of oracle pruning nowadays. Finally, given the evidence, we argue that the retraining stage in a pruning algorithm should be accounted for when developing pruning criteria.

cs.LG↗

arXiv2026

Leaky-Wave Antenna for Non-Invasive Continuous Glucose Monitoring

Non-invasive continuous glucose monitoring requires sensing techniques capable of delivering stable, accurate, and real-time quantification of glucose-induced dielectric variations. This paper presents a microwave-based methodology that employs a leaky-wave antenna (LWA) to launch localized traveling electromagnetic waves and estimate glucose concentration through phase measurements at a fixed operating frequency. The LWA architecture enhances electromagnetic coupling with the material under test through controlled energy leakage, enabling strong linearity and high phase sensitivity within the physiological permittivity range (relative permittivity 68-70). In vitro measurements using fetal bovine serum-glucose samples demonstrate a total phase shift of 8.51 degree across glucose concentrations from 50 to 255 mg per liter. Preliminary in vivo experiments conducted on two human subjects show strong correlations between measured S11 phase and reference glucose levels (r=0.939 and r=0.906), confirming the potential of the LWA-based sensing mechanism to track dynamic glycemic variations in a proof-of-concept setting. These results highlight the potential of the proposed approach for practical, continuous, and non-invasive glucose monitoring.

physics.app-ph↗

arXiv2026

EGI-BENCH: An Enterprise General Intelligence Benchmark Suite

Enterprise adoption of LLMs in agentic workflows is gated by evaluations that span a wide arc of abilities: CRM operations, heterogeneous-document retrieval, personal data understanding, multi-turn planning, tool use, function calling, long-horizon research, and long-context software engineering. These evaluations live today in separate repositories with divergent runners, dependencies, and scoring conventions, which makes cross-benchmark comparison difficult and obscures what a single aggregate number actually represents. We release EGI-BENCH, a unified LLM agent benchmark suite that consolidates eight enterprise facing agent benchmarks (CRMARENA, HERB, PERSONABENCH, USERBENCH, MCP-UNIVERSE, MCPEVAL, BENCHFORCE, LOCOBENCH) behind a single command-line interface with two runtime-budgeted execution modes. On top of the suite we define the Enterprise General Intelligence (EGI) Score, a six-dimensional framework (M1-M6) whose buckets are populated by fine-grained sub-metrics routed from each benchmark rather than by benchmark-level averages. We describe the suite, the sub-metric routing, and the quick/deep budgets, and report results for fourteen frontier and cost-efficient LLMs spanning the GPT, Claude, and Gemini families under quick mode. Top-line EGI ranges from 39.6 (GPT-4o mini) to 65.6 (gpt-5.5), and no single model dominates every capability axis: gpt-5.5 leads task outcome, planning, and retrieval grounding; Claude Sonnet 4.5 leads tool actioning and safety/abstention; and Claude Haiku 4.5 leads engineering quality. Code is available at https://github.com/SalesforceAIResearch/egi-bench.

cs.AI↗

arXiv2026

EdgeKernelBench: Unified LLM-Driven Kernel Generation for Edge Backends

Recent works show that LLMs can generate kernels with strong performance on server GPUs. Edge deployment presents distinct challenges because the hardware is heterogeneous, the software stacks are fragmented, and the resource budgets are tight. LLM kernel generation across diverse edge backends has yet to be studied systematically. We introduce EdgeKernelBench, which combines a unified IR deployment framework with a hierarchical benchmark for edge workloads. The unified IR captures reusable computation, scheduling, and memory decisions and connects to each target backend through a compiler interface. The benchmark spans progressively larger program scopes, from individual operators to compact model graphs, across representative edge application domains. We instantiate the design with TVM as a reference compiler and validate it across heterogeneous edge backends. With a single generation attempt, the best model achieves 64.5% Correct@1 but only 25.5% Fast@1.0. This gap shows that functional correctness does not consistently translate into performance gains. We therefore introduce the Layered Edge Kernel Agent (LEKA), which organizes optimization through a layered IR. On the agent evaluation subset, LEKA achieves 43.5% Fast@1.0, compared with 13.0% for direct generation and 26.1% for Claude Code. LEKA and Claude Code use the same base model and optimization budget. Together, EdgeKernelBench and LEKA support unified evaluation and layered optimization of kernels generated by LLMs across edge backends.

cs.LG↗

arXiv2026

Learning Personalized Memory Gating from Delayed One-Sided Feedback

Long-term memory enables language-model agents to personalize future interactions, but a finite memory bank cannot preserve every extracted fact. Existing memory gating policies are typically shared across users and therefore fail to account for user-specific patterns of future reuse. In this paper, we study memory gating at the fact level. We first propose a retrieval-based labeling strategy that labels each fact by whether it is later retrieved and incorporated into responses. We then formulate online memory gating as contextual apple tasting with delayed one-sided feedback, and propose Warm-Start UCB, which combines offline learning from existing data with online adaptation to the current user. Our theoretical analysis establishes a high-probability regret upper bound of $\tilde{O}(Tε\wedge d\sqrt{T})$ when the delay window is at most $O(\sqrt{T})$, where $T$ is the total number of rounds, $ε$ is the initialization error, $d$ is the dimension of the feature map, and $\tilde{O}(\cdot)$ hides constant and poly-logarithmic factors. The results of our experiment demonstrate that our method outperforms all baselines under different memory budgets and history lengths, establishing it as a promising personalized memory gating policy.

cs.LG↗

arXiv2026

Salesforce Koa: An Enterprise Language Model for Agentic Tool Use

We present Salesforce Koa, an enterprise language model built by post-training the open-weight Nemotron-3-Super-120B foundation model with reinforcement learning using Group Relative Policy Optimization (GRPO), and deployed in FP8 for production. Koa is trained only on public and synthetically generated data, and specialized for the agentic tool use that enterprise workflows demand: routing a request to the correct action, invoking the right tool with valid arguments, and completing multi-turn business tasks. The distinctive component of our pipeline is specification-driven task construction: declarative Agent Script specifications are expanded into persona-conditioned multi-turn environments whose rewards are grounded in successful tool use. Applied to enterprise CRM specifications, the same pipeline produces the in-domain training distribution on which Koa is specialized. On CRMAgentBench and the human-labeled production tool-calling set, Koa outperforms both its untuned open-weight base and GPT-4.1 and is competitive with the strongest frontier models. It reaches 87% Task Success Rate on CRMAgentBench (vs. GPT-4.1 at 82% and the base at 79%), is at or near the top of every metric on the human-labeled portion of an internal production benchmark, and preserves the base model's general capability on public benchmarks (Tau2Bench, BFCL). A controlled comparison with architecture and RL recipe held fixed shows that the additional in-domain RL stage improves argument accuracy and full tool-call success on the human-labeled enterprise benchmark.

cs.CL↗

arXiv2026

Universal Textual Teaching for LLMs

Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require training the Student parameters, thereby binding the distilled knowledge to a specific architecture and checkpoint. This implicit representation is difficult to interpret or reuse across models and limits KD for API-only or costly-to-train models. This paper studies knowledge transfer for large language models (LLMs). We introduce Universal Textual Teaching (UTT), a parameter-update-free framework that distills observed Teacher-Student knowledge gaps into a textual, interpretable, and reusable natural-language artifact called Primer. Specifically, UTT first identifies representative gap cases through paired evaluations, and iteratively updates the Primer via multi-role interactions: the Student attempts each task, the Prompter turns evaluation feedback into a teaching instruction, the Teacher provides a targeted demonstration, and the Synthesizer consolidates validated lessons. Empirically, on the challenging math (Omni-MATH-2) and code generation (KernelBench) tasks, extensive results confirm the effectiveness of the method: UTT remarkably raises the Student's accuracy from 9.4% to 48.6% and Fast1 accuracy from 9% to 35% on KernelBench, while increasing mathematical reasoning accuracy from 27.6% to 51.7%. UTT also performs better than representative prompt engineering and parameter-based KD methods. Of note, UTT is shown to be generalizable across different Teachers and Students: a Primer synthesized for one Teacher-Student pair can generalize to other Students that do not participate in the synthesis.

cs.AI↗

arXiv2026

SparseDecoding: Decoding-Aware Pruning for Accurate and Efficient LLM Inference

The memory-bound nature of the decoding stage of large language model (LLM) inference incurs significant latency. Layer-wise training-free network pruning approaches guided by the Hessian have been a prominent solution to this problem, as pruning reduces the number of nonzero parameters read from memory during decoding. Nevertheless, typical methods in this line compute the Hessian using pre-collected natural sequences, whereas the model is fed self-generated tokens during decoding, creating a distribution shift between the two sequences. The Hessian calculated on the natural sequence is different from that calculated on the generated sequence. We observe that this discrepancy causes the activation distribution during generation to deviate from that used for pruning, further hurting the pruned model performance. Moreover, most existing LLM pruning methods that bring actual speedup primarily target the sparse matrix-matrix (SpMM) multiplication, providing limited support for the sparse matrix-vector (SpMV) operations, which dominate decoding. To solve these problems, we introduce SparseDecoding, a principled decoding-aware pruning framework tailored for accurate and efficient LLM decoding. Specifically, at the algorithmic axis, SparseDecoding constructs calibration matrices from layer-wise activations collected during the dense-model autoregressive generation, excluding prefill, thereby aligning the pruning objective with the decoding activations. At the system axis, we develop an optimized N:M sparse matrix-vector kernel with bitmask indexing and fixed-step traversal. Substantial empirical results on representative LLMs (Llama-3.1-8B, Llama-3.3-70B, Qwen3-14B / 32B) demonstrate that our method consistently outperforms standard fixed-text calibration on the long-form generation benchmarks while achieving up to 1.48x end-to-end wall-clock decoding speedup on A100 GPUs.

cs.LG↗

arXiv2026

Boosting Large Language Models with Mask Fine-Tuning

The large language model (LLM) is typically integrated into the mainstream optimization protocol. However, it remains underexplored whether maintaining the model integrity is \textit{indispensable} for promising performance. In this work, we introduce Mask Fine-Tuning (MFT), a novel LLM fine-tuning paradigm demonstrating that carefully breaking the model's structural integrity can surprisingly improve performance without updating model weights. MFT learns and applies binary masks to well-optimized models, using the standard LLM fine-tuning objective as supervision. Based on fully fine-tuned models, MFT uses the same fine-tuning datasets to achieve consistent performance gains across domains and backbones (e.g., an average gain of 2.70/4.15 on IFEval with LLaMA2-7B/3.1-8B). Detailed ablation studies and analyses examine the proposed MFT from different perspectives, including the sparse ratio and the loss surface. Additionally, when deployed on well-trained models, MFT is compatible with other LLM optimization procedures to improve overall model performance. Furthermore, this study extends the masking operation beyond its conventional use in network pruning for model compression to encompass a broader range of model capabilities.

cs.CL↗

arXiv2026

MetaKernelBench: Measuring GPU Kernel Knowledge Transfer Beyond Code

Recent GPU kernel optimization agents retain what they learn in knowledge bases or as distilled skills. Kernel benchmarks score each attempt's implementation for correctness and speed but leave the reuse value of retained experience unmeasured. We introduce MetaKernelBench, which measures whether experience distilled from an attempt in one kernel domain-specific language (DSL) improves a fresh attempt at the same problem in another. Its 74 problems are fused subgraphs in six families, each posed as a pair of CuTe DSL and TIRx variants that differ only in the DSL. The agent first attempts each variant solo and is instructed to distill what it learns into a natural-language skill, which is transferred whether or not the source attempt passes verification. The skill is the only extra input to a skill-conditioned attempt by the same model in the other DSL. We compare each skill-conditioned attempt with the solo attempt on the same variant under matched per-attempt budgets, scoring correctness and end-to-end runtime. Across six models and both directions, paired lift over solo attempts ranges from -19% to +29%. Four models gain in both directions, yet regressions occur on 16% to 45% of problems in every model and direction. Outcomes follow the source attempt's result relative to the target's solo attempt rather than source success alone, improving in 71% of comparisons when the source stands above and regressing in 54% when it stands below. MetaKernelBench complements implementation-quality metrics by measuring same-problem cross-DSL kernel knowledge transfer.

cs.LG↗

arXiv2026

Rewarding Novel Deductions: Solver-guided Process Supervision for Logical Reasoning

Logical reasoning remains a major challenge for large language models (LLMs), particularly on structured problems that require precise constraint tracking, consistency preservation, and multi-step deduction. This challenge is especially acute for small-scale LLMs, which are more prone to producing inconsistent, redundant, or brittle reasoning trajectories. Existing approaches for improving logical reasoning largely optimize for final-answer correctness, providing only weak supervision over the intermediate reasoning process. In this work, we propose SPRING: (Solver-guided Process Rewards for Novel LogIcal ReasoNing Step Generation). SPRING uses SMT solver as a training-time verifier of intermediate reasoning steps to provide process-level supervision. It introduces the notion of a novel reasoning step, namely, a step that is logically valid, consistent with the evolving reasoning state, and not already implied by previously accepted non-contradictory deductions. Based on this solver-based assessment, it designs process rewards that encourage novel inferential progress while penalizing contradictory and uninformative reasoning steps. Evaluation across three logical reasoning benchmarks, ZebraLogic, AR-LSAT, and Knights and Knaves, and four LLMs shows that SPRING consistently outperforms base LLMs, outcome-only reward baselines, and Logic-LM. On ZebraLogic, SPRING improves puzzle accuracy by up to 49.71 and 15.43 points over the base LLM and strongest outcome-only baseline, respectively. On AR-LSAT, it improves overall accuracy by up to 64.93 and 12.14 points, respectively. On Knights and Knaves, SPRING achieves up to 93.14 puzzle accuracy and 96.05 person accuracy.

cs.CL↗

arXiv2026

Symmetry-enabled tunable square-lattice Hubbard models in $Γ$-valley moiré bilayers

Square-lattice moiré systems provide a promising route for quantum simulation of Hubbard physics, yet electrically tunable realizations of the hopping ratio $t'/t$ have so far been limited to specific valley configurations. A natural question is whether the same tunability can be extended to the more broadly occurring $Γ$-valley systems. Here we identify a symmetry-controlled route for realizing electrically tunable square-lattice Hubbard models in $Γ$-valley twisted homobilayers. At small twist angles, an emergent layer-exchange symmetry separates the low-energy states into flat bands localized on two nested square sublattices, suppressing inter-sublattice hopping. An interlayer displacement field breaks this symmetry and induces controllable hybridization between the sublattices, enabling continuous tuning of the $t'/t$ ratio over a wide range while preserving an effective single-band description. We further establish a formal correspondence between $Γ$- and M-valley moiré systems, revealing them as different symmetry limits of a unified framework for tunable square-lattice Hubbard models. Our results uncover a general symmetry principle underlying displacement-field tunability and extend electrically controllable square-lattice Hubbard physics to a broad family of $Γ$-valley materials.

cond-mat.mes-hall↗

arXiv2026

Beyond Normal References: Discriminative Few-Shot Anomaly Detection

This paper considers a practical few-shot anomaly detection (FSAD) setting, termed discriminative FSAD, where a limited number of both normal and anomalous examples are available as references during inference. Existing FSAD methods rely on normal-only references through normality matching, ignoring the discriminative clues in anomalous references, while directly fitting both references can overfit to the seen anomalies. We introduce IDEAL, an intrinsic deviation learning framework that leverages both reference types to learn intrinsic deviation patterns characterizing generalizable abnormality as deviations from normality. IDEAL decomposes the learning process into two novel components: 1) a Normal Variation Eraser to suppress nuisance normal variations that may lead to noisy deviations from normality, thereby highlighting anomaly-relevant deviation representations; 2) an Intrinsic Deviation Encoder to decompose these denoised deviation representations into intrinsic deviation vectors capturing the most discriminative orthogonal deviation directions. At inference, IDEAL scores query-to-normal deviations preserved after projection onto the learned intrinsic deviation vectors, enabling generalization for both seen and unseen anomalies. Extensive experiments on eight real-world datasets show that IDEAL generalizes effectively to unseen anomalies and consistently outperforms existing state-of-the-art FSAD methods. Code and data are available at \href{https://github.com/mala-lab/IDEAL}{https://github.com/mala-lab/IDEAL}.

cs.CV↗

arXiv2026

RamanPFN: learning from Raman spectral structure with a tabular foundation model

Raman spectroscopy enables label-free molecular characterization across materials science, analytical chemistry, biomedicine, and industrial process monitoring. However, machine learning for high-dimensional spectroscopy remains constrained by limited labelled data and a mismatch between the physical organization of spectra and feature-agnostic models. Channel coverage alone does not ensure that related bands share a common inference context. Here we present RamanPFN, a general-purpose spectral foundation framework that enables unified in-context inference through physics-guided spectral learning. It captures full-spectrum compositional covariation via Global Compositional Unmixing (GCU), which decomposes distributed, multi-band mixture signatures into shared non-negative latent bases. Simultaneously, it resolves local vibrational structure through Local Vibrational Subspace Encoding (LVSE), which preserves fine-grained peak morphology, intensity fluctuations, and peak shifts within contiguous spectral neighborhoods. Extensive evaluation across 74 diverse public Raman datasets covered 129 regression targets and was further extended to 21 classification tasks. RamanPFN achieved state-of-the-art performance across all reported aggregate metrics against 28 independently reproduced methods spanning chemometrics, spectral neural networks, deep tabular learners and tabular foundation models. RamanPFN establishes a physics-guided paradigm for scientific spectroscopy, enabling data-efficient predictive learning across diverse chemical systems.

cs.LG↗

arXiv2026

Standardized Dynamic Borrowing: Uniform Control of Information Borrowing

Dynamic borrowing can improve efficiency in clinical trials using external information, but existing methods may not offer transparent, prespecified control over borrowing. We introduce Standardized Dynamic Borrowing (SDB), which combines a source-informed component with a no-borrowing component and updates their relative odds using a Bayes factor standardized by its supremum over admissible target data. The posterior density under the no-borrowing component is obtained by normalizing the target-data likelihood. The prespecified source-component initial weight is a sharp uniform upper bound on its posterior weight, providing an interpretable borrowing constraint. When source and target models belong to the same one-parameter canonical exponential family, posterior expected local-information-ratio effective sample size equals target sample size plus borrowed source-prior effective sample size minus a nonnegative heterogeneity penalty. For normally distributed outcomes, borrowing depends on a standardized source-target conflict $Z$-score and decreases according to a standard normal kernel. We also jointly select the initial weight and target sample size to meet prespecified power and Type I error requirements. Normal and binomial illustrations demonstrate how SDB translates source-target compatibility into data-adaptive borrowing and how the calibration procedure prespecifies the source-component initial weight and target sample size under operating-characteristic constraints. These results provide a transparent framework for prespecifying, controlling, and interpreting dynamic borrowing.

stat.ME↗

arXiv2026

Random Attention: Rethinking KV Cache Eviction for Efficient Reasoning

Large language models achieve superior performance on tasks that require extended reasoning, but long chains of thought make the KV cache a severe memory bottleneck. Existing KV cache compression methods share one paradigm: score each cached token by some estimate of how much it will matter later, and keep the top-scoring ones. We show that the selection signal contributes almost nothing. Random Attention keeps the prompt and evicts uniformly at random within each attention head, computing no score at all; across four models and six reasoning tasks, it matches the strongest baseline in task performance while delivering 32-43% higher throughput than that method when deployed with vLLM. Controlled experiments explain this by showing that 1) the prompt is the fragile part of the cache, and most of the gap between selectors is just whether their selection signal happened to keep it; 2) the reasoning trace protects itself against eviction with redundancy at two levels, in the text (the model restates what it still needs as it works) and across attention heads (each keeps its own copy of the trace), so once the prompt is safe, a random draw retains enough copies of what the model still needs, and no score is required to pick them. Our code is publicly available at https://github.com/SalesforceAIResearch/Random-Attention.

cs.CL↗

arXiv2026

Critical-State RL: Diagnosing Trainable States for Multi-Turn Tool Use

Multi-turn tool-use failures can hinge on a single model call, yet reward variation alone does not reveal which call would benefit from training. When rewards depend on later interactions, their variation can reflect downstream randomness rather than differences between the current actions. We introduce Critical-State RL to identify trainable states in multi-turn interactions. Given task-defined candidate calls and local rewards, the method assesses whether each reward captures the action's effect on task success and whether improvement over a reference policy is possible. It then uses nested sampling to separate action-dependent reward variation from continuation noise and optimizes the policy at the selected states using contextual-bandit training. Experiments on the Berkeley Function Calling Leaderboard (BFCL) v4 compare training at diagnostic-selected states with training at alternative states. For missing-function tasks, the diagnostic selects the response after the tool becomes available; for missing-argument tasks, it selects the response before the missing argument is supplied. Training the selected responses improves performance, including about 14 percentage points on the missing-function task, while training the alternatives leaves performance flat or worse. We further apply the recipe across models and tasks, including logged repeat-call avoidance and memory management.

cs.LG↗

arXiv2026

GameASG-Bench: Benchmarking Autonomous Software Generation for Game Development

Autonomous software generation (ASG) aims to turn human requirements into executable applications, but delivering these applications does not necessarily establish that their interacting components satisfy the specified behavioral requirements. We introduce GameASG-Bench, a benchmark that makes behavioral testability part of the generation task for game development. Our design declares an evaluation interface specification before generation, fixing legal starting scenarios, player-level actions, stable snapshots, rejection behavior, and invariants while leaving private implementations open. Concretely, we include: (i) static L1 checks that assess source-level compliance; and (ii) browser-executed L2 checks that combine semantic observations with real input and runtime evidence. We implement this protocol as 47 browser-native game-generation tasks spanning 12 primary genres and both 2D and 3D interaction, each with executable checks and an independently verified reference implementation. Our experiments answer four key questions about end-to-end agent performance, tool access and nominal turn budget, reasoning effort, and harness choice. Across nine agent stacks, the highest observed mean L2 check pass rate is 93.2%, yet the highest observed strict task success rate, requiring all L1 and applicable L2 prerequisite and core requirement checks, is only 55.3% (26/47 tasks). For DeepSeek-V4-Flash, full tool access and larger nominal turn budgets yield more strict task successes, while the strict task success rate is not monotonic in reasoning effort. Both tested harnesses achieve 18 strict task successes, but only ten tasks succeed under both. These results expose task-level compliance gaps that high average check pass rates actually obscure.

cs.AI↗