跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+758新增论文
+578其中 AI 相关
+94已有论文修订
+95新增 AI 导读

arXiv 最近完成扫描:2026-11-17 04:58 · 全站最近入库:2026-11-17 04:58各信源进度 →

统计口径与数据新鲜度

统计 2026-11-16 09:09 — 2026-11-17 09:09(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-17 04:58最近发现新增版本:2026-11-17 04:58库内论文最新更新时间:2026-11-14 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

24,106 已收录论文20,158 AI 相关7,049 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 18 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-17 04:57(北京时间)
新论文 · v1arXiv:2611.17879

From Local Prediction Error to Final Outputs: A Theory of Expert Sharing in Multi-Step Reasoning

Tiantong Wang, Zhiren Gong, Yixin Wang, Tiantong Wu, Wei Yang Bryan Lim

原文摘要

Mixture-of-experts (MoE) models can share experts across layers or computational steps, pooling training information from different contexts. This can reduce local prediction error, but multi-step systems are ultimately evaluated by their final outputs. We study when these local gains survive composition. In specified linear regression models, we derive exact sharing and merging criteria by balancing variance savings against bias from sharing across roles with different target behavior. Under a common fixed linear propagation map, we characterize expected squared final-output risk, retaining error directions and correlations. We construct cases where sharing lowers average local error yet increases final-output error. Within this setting, we also identify when orthogonal full pooling with isotropic noise preserves every local improvement at the final output. For general routed computations, we derive conditional end-to-end error bounds that account for distribution shifts, errors in surrounding components, and changes in expert selection. Separately, we show that correct predictions and route labels alone do not establish that individual experts implement the attributed operations. Controlled experiments realize a strict local-to-final risk reversal. In low-data symbolic settings, trained Transformers exhibit statistically supported teacher-forced one-step gains without statistically resolved free-running final-output gains. Separate intervention tests distinguish successful multi-layer actions from single-expert attribution and generalization to unseen transitions. Together, these results distinguish when sharing improves local estimation, when those gains survive to the final output, and what evidence is needed to attribute reusable operations to individual experts.

cs.LG · cs.AI

版本更新 · v3arXiv:2605.11490

Adaptive Calibration in Non-Stationary Environments

Junyan Liu, Haipeng Luo, Lillian J. Ratliff

原文摘要

Making calibrated online predictions is a central challenge in modern AI systems. Much of the existing literature focuses on fully adversarial environments where outcomes may be arbitrary, leading to conservative algorithms that can perform suboptimally in more benign settings, such as when outcomes are nearly stationary. This gap raises a natural question: can we design online prediction algorithms whose calibration error automatically adapts to the degree of non-stationarity in the environment, smoothly interpolating between i.i.d. and adversarial regimes? We answer this question in the affirmative and develop a suite of algorithms that achieve adaptive calibration guarantees under multiple calibration measures. Specifically, with $T$ being the number of rounds, $W$ being the unknown number of i.i.d. segments of the environment, and $C\in[0,T]$ being another unknown non-stationary measure defined as the minimal $\ell_1$ deviation of the mean outcomes, our algorithms attain $\widetilde{\mathcal{O}}(\min\{\sqrt{T}+(TC)^{\frac{1}{3}}, \sqrt{WT}\})$ for $\ell_1$ calibration error and $\widetilde{\mathcal{O}}(\min\{(1+C)^{\frac{1}{3}}, W\})$ for both $\ell_2$ and pseudo KL calibration error. These bounds match the optimal rates in the stationary case ($C=0$ and $W=1$) and recover known guarantees in the fully adversarial regime ($C, W=Ω(T)$). Our approach builds on and extends prior work (Hu et al., 2026; Luo et al., 2025), introducing an epoch-based scheduling together with a novel non-uniform partition of the prediction space that allocates finer resolution near the underlying ground truth.

cs.LG · stat.ML

版本更新 · v2arXiv:2611.10655

STABLE: A Continual Learning Optimizer with Adaptive Drift Control

Xiaoyan Li, Lanpei Li, Massimo Coppola, Vincenzo Lomonaco, Antonio Carta

原文摘要

Continual learning suffers from catastrophic forgetting, leading to the loss of past knowledge and unstable adaptation across task streams. Existing approaches typically address this issue through replay buffers, architectural changes, or additional regularization terms. We argue that forgetting is a consequence of unstable learning dynamics, which may benefit from an optimizer-level solution. The main challenge is that learning dynamics are time-dependent, as demonstrated by empirical phenomena such as the stability gap. First, we formally prove that, for distance-based classifiers, bounded parameter drift guarantees bounded margin erosion under a local score-stability condition. Based on this insight, we propose STABLE, a primal-dual optimizer with adaptive drift control. STABLE enforces a drift budget by increasing the dual variable when the budget is violated and decreasing it otherwise. This results in high stability without penalizing plasticity. Experiments on long-horizon unimodal streams and CLIP-based multimodal streams show that vanilla PEFT baselines trained with STABLE match or outperform specialized continual learning methods. In the multimodal setting, STABLE further preserves zero-shot transfer, matching or exceeding dedicated multimodal continual learning methods. Code is available at https://github.com/atlas-luiss/STABLE.

cs.LG · cs.AI · cs.CV

新论文 · v1arXiv:2611.14097

Stylization as Geodesic Navigation on Diffusion Manifolds

Yan Li, Zehao Li, Xinlong Sun, Fuchen Long, Xi Chen, Wenhan Luo, Yike Guo

原文摘要

Facial stylization asks for a delicate transformation: the image should look unmistakably artistic, yet the person should remain unmistakably the same. Existing training-free and instruction-driven methods often fail to expose such control, either drifting away from identity or producing only weak stylization. We introduce Stylus, a training-free framework that treats facial style transfer as navigation on the curved latent manifold of a pretrained diffusion model. Rather than interpolating in flat Euclidean space, \method refines the interpolated content--style point by lowering a ratio-weighted discrete geodesic energy under the pullback metric, made tractable by a rank-one power-iteration approximation of the Jacobian. This principle drives two complementary controls: content--style latent interpolation for smooth semantic fusion, and self-attention query optimization for spatially adaptive style injection. The resulting dual-control mechanism provides continuous, fine-grained control over stylization strength while preserving identity-critical facial structure, all without additional training. Experiments show that Stylus achieves the strongest identity and content preservation at competitive stylization strength.

cs.CV · cs.LG

版本更新 · v2arXiv:2605.12541

PPG2ECG: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation

Xiaoda Wang, Minxiao Wang, Kaiqiao Han, Defu Cao, Ching Chang, Yidan Shi, Runze Yan, Xiao Luo 等 13 位作者

原文摘要

Electrocardiography (ECG) is the clinical standard for cardiac assessment but requires dedicated hardware that does not scale to daily-life monitoring. Photoplethysmography (PPG) is ubiquitous in wearables but lacks ECG-specific diagnostic morphology and is corrupted by motion and sensor noise. PPG-to-ECG generation aims to bridge this gap by recovering electrical morphology and timing from peripheral pulse signals. However, existing methods largely rely on statistical alignment and data-driven generation. They fail to explicitly structure the latent space around physiology-aware electro-hemodynamic factors and lack constraints from forward physiological dynamics. To address these challenges, we propose PPG2ECG, a physiology-guided latent rectified flow framework. PPG2ECG introduces an electro-hemodynamic simulator that co-models ECG and PPG through shared cardiac phase dynamics. Guided by this simulator, a Physiology-Aware AutoEncoder learns a structured electro-hemodynamic latent space. Then we integrate this simulator guidance into a PPG-conditioned latent rectified flow, enforcing ECG-side morphology consistency and ECG-to-PPG forward hemodynamic consistency during generative transport. Experiments on the large-scale MC-MED dataset demonstrate that PPG2ECG significantly improves PPG-to-ECG generation and downstream cardiovascular disease classification, proving its ability to generate ECGs that are both signal-faithful and physiologically plausible under the ECG-to-PPG hemodynamic pathway

eess.SP · cs.AI · cs.LG

新论文 · v1arXiv:2611.11102

Task presentation and execution paths shape LLM mathematical performance

Min Huang, Mingyan Liu

原文摘要

Mathematical performance depends on how a language model receives information, uses generated text and submits an answer. We study these dependencies through paired comparisons that separate submission requirements, draft reuse and candidate availability. On 65 development problems, an instruction permitting continued derivation produced no valid final answers in 390 short-output calls; a single-line submission instruction produced 390 format-valid outputs, including 389 parseable answers. Under the fixed submission protocol, 198 problems with failed initial submissions were answered correctly in 9.09p of cases with the question alone, 33.84p with its draft, 46.46p with candidates, and 59.60p with both. Drafts added 13.13 percentage points with candidates fixed, and candidates added 25.76 points with drafts fixed; both effects were significant after Holm correction and were subadditive on the accuracy scale. Retaining a trailing excerpt outperformed retaining a leading excerpt of similar average length by 10.10 points Holm-adjusted p 0.027. A second completion model showed the same directions of input effects on the fixed materials. In a cohort with parseable initial submissions, candidates added no correct answers when the previous draft was supplied, but added 27 points when it was absent. These results establish a conditional value of generated text: a failed submission can leave useful material, whose contribution depends on what else is supplied and what is retained.

cs.LG · cs.CL · cs.SE

新论文 · v1arXiv:2611.10936

Can small transformers learn the language of Coxeter groups?

J. Elisenda Grigsby, Yiyao Huang, Arushi Katyal, Ryan Lin, Pranav Ramakrishnan, Weiya Wang

原文摘要

The field of mechanistic interpretability is well-suited to collaborative projects between mathematicians and AI researchers. This work describes one such project. We trained a small transformer to learn the language of the Coxeter group affine A2, best known as the symmetries of the tiling of the Euclidean plane by equilateral triangles. We then formulated and proved a word-theoretic conjecture about affine A2 after observing the attention patterns of the trained model.

cs.CL · cs.LG · math.GR · math.RT

版本更新 · v4arXiv:2605.01973

Learn-To-Learn on Arbitrary Textual Conditioning: A Hypernetwork-Driven Meta-Gated LLM

Luo Ji, Qi Qin, Ningyuan Xi, Teng Chen, Qingqing Gu, Hongyan Li

原文摘要

Conventional LLMs may suffer from corpus heterogeneity and subtle condition changes. While finetuning can create the catastrophe forgetting issue, application of meta-learning on LLMs is also limited due to its complexity and scalability. In this paper, we activate the meta-signal of $β$ within the SwiGLU blocks, resulting in a meta-gating mechanism that adaptively adjusts the nonlinearity of FFN. A hypernetwork is employed which dynamically produces $β$ on textual conditions, providing meta-controllability on LLMs. By testing on different condition types such as task, domain, persona, and style, our method outperforms finetuning and meta-learning baselines, and can generalize reasonably on unseen tasks, condition types, or instructions. Our code can be found in https://github.com/AaronJi/MeGan.

cs.CL · cs.LG

新论文 · v1arXiv:2611.10536

In Search of Lost Time: Temporal Competence in Large Language Models

Ryan Liu, David Lee, Icey Siyi Ai, Marwa Abdulhai, Rachit Dubey, Thomas L. Griffiths

原文摘要

Today's large language models (LLMs) increasingly interact with time: AI agents schedule tasks, coordinate with people and other AI systems, and even decide how long to think. Yet LLMs only ever encounter time secondhand---in descriptions rather than experience. Does this hinder their reasoning and decisions that involve time? Our paper traces this question from a model's knowledge to its actions through four successive evaluations. We find that models know the rules of time, which can be read, but struggle to predict durations, which must be lived; while they are highly accurate (92.0-97.5\%) on a collection of questions built from the theorems of various time-related literatures, they underestimate their own completion times by up to 99.5\% and fail to identify which tasks will take them longer to run. And while they are better at estimating durations of human tasks, their duration judgments change based on how an event is worded, magnifying similar human biases by up to 1.3$\times$. Moving from estimation to action, we find that some LLMs disregard time costs completely when answering user queries, ignoring a 1,200-fold increase in wait time entirely. Together, our experiments reveal new challenges posed to LLMs as they navigate a dimension they cannot directly perceive, and highlight how their blind spots can even be simple cognitive abilities that humans take for granted.

cs.AI · cs.CL · cs.LG

新论文 · v1arXiv:2611.07646

WaSP: Wasserstein Data Synthesis Pipeline for Solver-Free Operator Learning

Ryan Liu, Chuwei Wang, Ricardo Baptista, Andrew Stuart, Anima Anandkumar

原文摘要

Neural operators offer fast surrogates for solving PDEs, but training them typically requires expensive input-solution pairs. Two solver-free alternatives include physics-informed learning and training on manufactured data. Physics-informed learning approximates solutions by minimizing PDE residuals. While the residual norm must be strong enough to control solution error, stronger norms increase mini-batch variance and hinder optimization. Manufactured data constructs input-solution pairs for learning by applying the PDE operator to artificial solutions. However, the resulting training distribution can be mutually singular with the target distribution, yielding poor accuracy on the target distribution. We introduce the Wasserstein Data Synthesis Pipeline (WaSP), which trains on manufactured input-solution pairs while gradually mitigating the distribution shift. WaSP synthesizes new pairs by applying the learned operator to a path of intermediate distributions converging to the target input distribution. Training proceeds in stages along this path. At each stage, the previous surrogate generates solutions, and the exact PDE operator derives corresponding inputs, yielding training pairs for the next surrogate. Theoretically, we prove a stepwise error-propagation bound for the distribution shift introduced during synthesis. Empirically, WaSP reduces solution error by $3.9\times$ to $9.5\times$ compared to existing solver-free methods on PDEs such as Allen-Cahn, Kuramoto-Sivashinsky, and Ginzburg-Landau, and better reproduces the two chaotic systems' statistical observables.

cs.LG · cs.AI

版本更新 · v2arXiv:2609.03079

LeanStream: A Speculate-and-Refine Streaming Framework for Efficient on-Device LLM Inference

Renyuan Liu, Yuyang Leng, Kaiyan Liu, Yuzhou Zhong, Shaohan Hu, Chun-Fu, Chen, Peijun Zhao 等 10 位作者

原文摘要

On-device LLM inference is attractive for privacy and responsiveness, but remains challenging on mobile and embedded devices because model weights far exceed available DRAM. Prior systems exploit activation sparsity and offload weights to SSD or flash storage, but face a fundamental systems trade-off: accurate sparse execution decisions require the latest context, whereas efficient computation-I/O overlap requires early prediction. As a result, existing designs either serialize execution or incur redundant weight fetches, extra computation, and large cache overheads. We present LeanStream, a streaming speculate-and-refine framework for efficient on-device LLM inference. LeanStream progressively refines computation, loading, and cache-retention priorities using partial GPU results, enabling fine-grained overlap between GPU execution and storage I/O. We implement LeanStream on both mobile and embedded platforms. Compared with prior on-device LLM inference systems, LeanStream reduces memory usage by 4.8$\times$ to 7.5$\times$ at the best throughput achieved by prior work, while further improving token generation throughput by 1.6$\times$ to 2.1$\times$. Source code is available at https://github.com/RenyuanLIU/LeanStream.

cs.LG

新论文 · v1arXiv:2611.06606

Verifier-Guided Self-Consistency: Test-Time Reasoning with Imperfect Generators and Verifiers

Ekaterina Fadeeva, Xueyan Li, Mrinmaya Sachan, Maxim Panov

原文摘要

Large language models can solve many multi-step reasoning problems, but their reasoning remains brittle: plausible intermediate steps often lead to incorrect answers. Improving reasoning at test time requires balancing two imperfect signals: the generator, which provides diversity, and the verifier, which guides correctness. Existing approaches largely favor one side of this tradeoff. Verifier-weighted voting relies on trajectories sampled from the generator and cannot steer generation toward correct ones. By contrast, beam-style verifier-guided search can steer generation, but may over-concentrate on narrow trajectory sets and amplify verifier errors. We propose Verifier-Guided Self-Consistency (VGSC), a stochastic decoding framework that combines step-level correctness and future-outcome estimates to guide reasoning during generation. VGSC repeatedly proposes candidate next steps from the base language model and uses both signals in an acceptance/resampling procedure, while maintaining multiple trajectories for final answer-level self-consistency. We provide a probabilistic interpretation of VGSC as approximate sampling from a verifier-tilted trajectory distribution. Across multiple reasoning benchmarks, generator models, and verifier families, VGSC achieves higher average accuracy than representative baselines.

cs.LG

版本更新 · v2arXiv:2604.27613

AMGenC: Generating Charge Balanced Amorphous Materials

Yan Lin, Jilin Hu, N. M. Anoop Krishnan, Morten M. Smedskjaer

原文摘要

Amorphous (disordered) materials require simulation cells with at least hundreds to thousands of atoms to accurately capture their diverse local atomic environments, compared to crystalline materials that can be described by unit cells containing a few to hundreds of atoms. To advance the design of amorphous materials with desired properties and facilitate the exploration of their vast design space, generative inverse design has emerged as a promising approach. It aims to directly output materials with properties closely aligned with the desired ones using probabilistic generative models conditioned on desired properties, which can be more resource efficient than the traditional trial-and-error approach. However, due to the inherent stochasticity of probabilistic generative models, when element assignments are unconstrained, a large portion of generated materials may be charge unbalanced, and existing methods either fix the composition in advance or discard charge unbalanced samples after generation. In this work, we propose AMGenC, a new generative inverse design method for amorphous materials that can guarantee the generation of formally charge balanced samples, with minimal additional computational overhead and without sacrificing inverse design accuracy. AMGenC achieves this through an element noise that gives the generation process a starting point centered around charge balance, and the combination of a per-step soft projection and a final discrete projection for steering the elements toward exact charge balance throughout the generation. We perform extensive experiments on two amorphous materials datasets. Experimental results provide evidence that AMGenC achieves its design goal.

cs.LG

新论文 · v1arXiv:2611.03806

BEND 2.0: Benchmarking the Capabilities Needed to Design DNA

Andreas Bjerregaard, Rachael DeVries, Dionysia Danai Brilli, Panagiotis Antoniadis, Yan Li, Anders Krogh, Wouter Boomsma, Ole Winther 等 9 位作者

原文摘要

Genomic foundation models are increasingly used to propose DNA sequences and edits. But successful design requires capabilities that current benchmarks do not measure. Here we introduce BEND 2.0 (BENchmark for DNA language models), which tests how well models predict the effects of DNA changes and whether these predictions can guide sequence design. Across eight tasks built on five experimental datasets, we find that model choice and how representations are accessed both affect performance. Linear readouts fitted to frozen representations can match or exceed models' native outputs, and functional post-training substantially improves the models' capabilities. Linear readouts transfer well across experimental assays and genes, suggesting that models capture functional patterns beyond the data used to fit individual predictors. At the same time, strong prediction of average effects can conceal weaker prediction of context-specific effects, and accurate prediction of combined mutation effects does not imply accurate prediction of their interactions. In design tasks, model predictions can reduce the number of candidates needed to find a desired response, but still leave substantial room for better design choices, particularly for context-specific and multi-cell objectives. Our benchmark reveals that genomic models now display signs of general utility but have yet to saturate on performance in even simple tasks.

q-bio.GN · cs.LG · q-bio.QM

新论文 · v1arXiv:2611.01649

Near-Optimal $\sqrt{K}$-Regret for Online Learning in Uninformed Markov Games

Junyan Liu, Haipeng Luo, Lillian J. Ratliff, Zihan Zhang

原文摘要

We study online learning in two-player uninformed Markov games, where the opponent may follow an arbitrary, unknown policy in each episode and the learner does not observe the opponent's actions. [TWYS21] showed that achieving sublinear external regret requires an exponential dependence on the episode length $H$. Motivated by this limitation, they studied the weaker notion of Nash-value regret and obtained an $\widetilde{O}(K^{2/3})$ regret bound, where $K$ is the number of episodes. More recently, [LLZR26] introduced a stronger notion called empirical Nash-value regret and achieved an $\widetilde{O}(\sqrt{K})$ bound when the opponent's non-stationarity is moderate, including the case where the opponent follows a fixed policy. However, whether the $\sqrt{K}$ rate is achievable in the worst case remained open. In this paper, we resolve this open question by developing an algorithm that achieves $\widetilde{O}(H^{3/2}S\sqrt{AK})$ empirical Nash-value regret against arbitrary opponent policies, where $S$ is the number of states and $A$ is the number of learner's actions. Specifically, our algorithm combines global and local planning by maintaining a collection of candidate predictions of future losses, each associated with a policy. Global planning learns which candidates to favor across episodes, while local planning combines their policies differently at each state to make decisions. A key ingredient is to coordinate these two components to control the discrepancy between local loss estimates and the total loss of the executed policy, using only the learner's observed trajectories.

cs.LG · stat.ML

新论文 · v1arXiv:2610.39927

Learning Calibrated Confidence for Multi-Agent Collaboration

Chenxiao Fan, Chongming Gao, Gangyi Zhang, Yaxin Gong, Leyang Shen, Jiakai Wang, Dong Wang, Yang Liu 等 12 位作者

原文摘要

Complex information seeking on the Web often requires collecting evidence from multiple documents and reasoning across them, as in multi-hop question answering. Multi-agent systems address such tasks by dividing work among specialized agents, where upstream agents produce intermediate answers and supporting evidence that downstream agents build upon. However, upstream answers vary in correctness, and most systems provide no quantitative signal about it, leaving the downstream agent to judge whether to trust or verify what it receives based solely on the textual evidence. In this paper, we propose MASCAL. Each agent learns to produce a correctness probability for its answer by jointly optimizing task performance and confidence calibration in reinforcement learning; the upstream agent passes this probability at handoff, giving the downstream agent a quantitative basis for its decisions, and the final answer also carries a calibrated probability, quantifying the reliability of the system output. To guide the downstream agent in using this signal during training, we introduce confidence-aware weighting, which increases the learning weight at positions responsive to the upstream probability, so that the agent learns more from decisions where it actively uses the signal. Across three collaboration architectures spanning Web retrieval and reasoning tasks, MASCAL improves task accuracy and calibration quality for both agents, and generalizes to deeper collaboration chains and different model scales. Intervention analysis confirms that upstream confidence affects downstream decision behavior, and the learned final confidence supports selective answering and targeted rechecking.

cs.LG · cs.AI · cs.MA

版本更新 · v2arXiv:2503.13383

An Embarrassingly Simple Data Selection Strategy for Million-Scale Multi-Modal Instruction-Tuning

Mengyao Lyu, Yan Li, Huasong Zhong, Hui Chen, Jungong Han, Guiguang Ding, Zhenheng Yang

原文摘要

Recent research demonstrates that pretrained large language models (LLMs) require minimal but carefully curated data for instruction-tuning. Building on LLMs, multi-modal models further contend with far larger data volumes, heterogeneous sources, and cross-modal alignment, amplifying both the necessity and the complexity of data selection. However, existing methods rely on deeply extracted embeddings, gradients or narrow quality scores that are time-consuming to compute and limited in expressiveness, frequently failing to surpass a simple random-sampling baseline. To overcome these limitations, we step back from complex embedding spaces and target the fundamental objective of instruction tuning: how and what knowledge should be elicited from the pretrained model and exposed to users. We leverage interaction \underline{S}tyles as a proxy for structural diversity, and decompose data quality into capability-specific \underline{S}cores, yielding an explicit yet \underline{R}ich (SSR) profile for each sample. We further cast the selection as submodular coverage maximization over the capability--style grid, providing an approximation guarantee. SSR consistently delivers superior performance across two distinct architecture stacks, LLaVA-1.5 (Vicuna + CLIP) and LLaVA-OneVision (Qwen2 + SigLIP), two large-scale data pools, six data budget settings, and 14 multi-modal benchmarks. To our knowledge, SSR is the first data selection method validated at million-scale, achieving 99.1% of full-data performance using only 30% of a 2.6M pool. Furthermore, the SSR profile models and selected subsets transfer training-free across data pools and model architectures with capability extensibility.

cs.CV · cs.AI · cs.CL · cs.LG

版本更新 · v4arXiv:2512.24555

From Perception to Punchline: Empowering VLM with the Art of In-the-wild Meme

Xueyan Li, Yingyi Xue, Mengjie Jiang, Qingzi Zhu, Yazhe Niu

原文摘要

Generating humorous memes is a challenging multimodal task that moves beyond direct image-to-caption supervision. It requires a nuanced reasoning over visual content, contextual cues, and subjective humor. To bridge this gap between visual perception and humorous punchline creation, we propose HUMOR}, a novel framework that guides VLMs through hierarchical reasoning and aligns them with group-wise human preferences. First, HUMOR employs a hierarchical, multi-path Chain-of-Thought (CoT): the model begins by identifying a template-level intent, then explores diverse reasoning paths under different contexts, and finally anchors onto a high-quality, context-specific path. This CoT supervision, which traces back from ground-truth captions, enhances reasoning diversity. We further analyze that this multi-path exploration with anchoring maintains a high expected humor quality, under the practical condition that high-quality paths retain significant probability mass. Second, to capture subjective humor, we train a pairwise reward model that operates within groups of memes sharing the same template. Following established theory, this approach ensures a consistent and robust proxy for human preference, even with subjective and noisy labels. The reward model then enables a group-wise reinforcement learning optimization, guaranteeing providing a theoretical guarantee for monotonic improvement within the trust region. Extensive experiments show that HUMOR empowers various VLMs with superior reasoning diversity, more reliable preference alignment, and higher overall meme quality. Beyond memes, our work presents a general training paradigm for open-ended, human-aligned multimodal generation, where success is guided by comparative judgment within coherent output group.

cs.LG