论文检索

Optimization

Liuyuan Jiang, Quan Xiao, Lisha Chen, Tianyi Chen

Bilevel optimization, a hierarchical optimization paradigm, has gained significant attention in a wide range of practical applications, notably in the fine-tuning of generative models. However, due to the nested problem structure, most existing algorithms require either the Hessian vector calculation or the nested loop updates, which are computationally inefficient in large language model (LLM) fine-tuning. In this paper, building upon the fully first-order penalty-based approach, we propose an efficient value function-free (\textsf{PBGD-Free}) algorithm that eliminates the loop of solving the lower-level problem and admits fully single-loop updates. Inspired by the landscape analysis of representation learning-based LLM fine-tuning problem, we propose a relaxed flatness condition for the upper-level function and prove the convergence of the proposed value-function-free algorithm. We test the performance of the proposed algorithm in various applications and demonstrate its superior computational efficiency over the state-of-the-art bilevel methods.

Applications · Robotics

Kefei Zhu, Fengshuo Bai, YuanHao Xiang, Yishuai Cai, Xinglin Chen, Ruochong Li, Xingtao Wang, Hao Dong, Yaodong Yang, Xiaopeng Fan 等

Dexterous manipulation is critical for advancing robot capabilities in real-world applications, yet diverse and high-quality datasets remain scarce. Existing data collection methods either rely on human teleoperation or require significant human engineering, or generate data with limited diversity, which restricts their scalability and generalization. In this paper, we introduce DexFlyWheel, a scalable data generation framework that employs a self-improving cycle to continuously enrich data diversity. Starting from efficient seed demonstrations warmup, DexFlyWheel expands the dataset through iterative cycles. Each cycle follows a closed-loop pipeline that integrates Imitation Learning (IL), residual Reinforcement Learning (RL), rollout trajectory collection, and data augmentation. Specifically, IL extracts human-like behaviors from demonstrations, and residual RL enhances policy generalization. The learned policy is then used to generate trajectories in simulation, which are further augmented across diverse environments and spatial configurations before being fed back into the next cycle. Over successive iterations, a self-improving data flywheel effect emerges, producing datasets that cover diverse scenarios and thereby scaling policy performance. Experimental results demonstrate that DexFlyWheel generates over 2,000 diverse demonstrations across four challenging tasks. Policies trained on our dataset achieve an average success rate of 81.9\% on the challenge test sets and successfully transfer to the real world through digital twin, achieving a 78.3\% success rate on dual-arm lift tasks.

Computer Vision · Everything Else

Fangtong Sun, Congyu Li, Ke Yang, Yuchen Pan, Hanwen Yu, Xichuan Zhang, Yiying Li

Low-light vision remains a fundamental challenge in computer vision due to severe illumination degradation, which significantly affects the performance of downstream tasks such as detection and segmentation. While recent state-of-the-art methods have improved performance through invariant feature learning modules, they still fall short due to incomplete modeling of low-light conditions. Therefore, we revisit low-light image formation and extend the classical Lambertian model to better characterize low-light conditions. By shifting our analysis to the frequency domain, we theoretically prove that the frequency-domain channel ratio can be leveraged to extract illumination-invariant features via a structured filtering process. We then propose a novel and end-to-end trainable module named \textbf{F}requency-domain \textbf{R}adial \textbf{B}asis \textbf{Net}work (\textbf{FRBNet}), which integrates the frequency-domain channel ratio operation with a learnable frequency domain filter for the overall illumination-invariant feature enhancement. As a plug-and-play module, FRBNet can be integrated into existing networks for low-light downstream tasks without modifying loss functions. Extensive experiments across various downstream tasks demonstrate that FRBNet achieves superior performance, including +2.2 mAP for dark object detection and +2.9 mIoU for nighttime segmentation. Code is available at: \url{https://github.com/Sing-Forevet/FRBNet}.

Applications · Language, Speech and Dialog

Yuhan Yao, Feifei Kou, Lei Shi, Xiao yang, Zhongbao Zhang, Suguo Zhu, Jiwei Zhang, Lirong Qiu, LI Haisheng

The advancements in generative models and the real-world attack of machine-generated text(MGT) create a demand for more robust detection methods. The existing MGT detection methods for adversarial environments primarily consist of manually designed statistical-based methods and fine-tuned classifier-based approaches. Statistical-based methods extract intrinsic features but suffer from rigid decision boundaries vulnerable to adaptive attacks, while fine-tuned classifiers achieve outstanding performance at the cost of overfitting to superficial textual feature. We argue that the key to detection in current adversarial environments lies in how to extract intrinsic invariant features and ensure that the classifier possesses dynamic adaptability. In that case, we propose OSTAR, a novel MGT detection framework designed for adversarial environments which composed of a statistical enhanced classifier and a Multi-Faceted Contrastive Learning(MFCL). In the classifier aspect, our Multi-Dimensional Statistical Profiling (MDSP) module extracts intrinsic difference between human and machine texts, complementing classifiers with useful stable features. In the model optimization aspect, the MFCL strategy enhances robustness by contrasting feature variations before and after text attacks, jointly optimizing statistical feature mapping and baseline pre-trained models. Experimental results on three public datasets under various adversarial scenarios demonstrate that our framework outperforms existing MGT detection methods, achieving state-of-the-art performance and robust against attacks.The code is available at https://github.com/BUPT-SN/OSTAR.

Applications · Everything Else

Yilun Zhao, Kaiyan Zhang, Tiansheng Hu, Sihong Wu, Ronan Le Bras, Yixin Liu, Robert Tang, Joseph Chee Chang, Jesse Dodge, Jonathan Bragg 等

We present SciArena, an open and collaborative platform for evaluating foundation models on scientific literature-grounded tasks. Unlike traditional benchmarks for scientific literature understanding and synthesis, SciArena engages the research community directly, following the Chatbot Arena evaluation approach of community voting on model comparisons.By leveraging collective intelligence, SciArena offers a community-driven evaluation of model performance on open-ended scientific tasks that demand literature-grounded, long-form responses.The platform currently supports 44 open-source and proprietary foundation models and has collected over 19,000 votes from human researchers across diverse scientific domains. Our analysis of the data collected so far confirms its high quality.We discuss the results and insights based on the model ranking leaderboard.To further promote research in building model-based automated evaluation systems for literature tasks, we release SciArena-Eval, a meta-evaluation benchmark based on our collected preference data. The benchmark measures the accuracy of models in judging answer quality by comparing their pairwise assessments with human votes. Our experiments highlight the benchmark’s challenges and emphasize the need for more reliable automated evaluation methods.

Theory · Learning Theory

Vaggos Chatziafratis, Ishani Karmarkar, Yingxi Li, Ellen Vitercik

Data-driven algorithm selection is a powerful approach for choosing effective heuristics for computational problems. It operates by evaluating a set of candidate algorithms on a collection of representative training instances and selecting the one with the best empirical performance. However, running each algorithm on every training instance is computationally expensive, making scalability a central challenge. In practice, a common workaround is to evaluate algorithms on smaller proxy instances derived from the original inputs. However, this practice has remained largely ad hoc and lacked theoretical grounding. We provide the first theoretical foundations for this practice by formalizing the notion of size generalization: predicting an algorithm's performance on a large instance by evaluating it on a smaller, representative instance, subsampled from the original instance. We provide size generalization guarantees for three widely used clustering algorithms (single-linkage, k-means++, and Gonzalez's k-centers heuristic) and two canonical max-cut algorithms (Goemans-Williamson and Greedy). We characterize the subsample size sufficient to ensure that performance on the subsample reflects performance on the full instance, and our experiments support these findings.

Applications · Physics

Jiyan Qiu, Lyulin Kuang, Guan Wang, Yichen Xu, Leiyao Cui, Shaotong Fu, Yixin Zhu, Rita Zhang

Vehicle aerodynamics optimization has become critical for automotive electrification, where drag reduction directly determines electric vehicle range and energy efficiency. Traditional approaches face an intractable trade-off: computationally expensive Computational Fluid Dynamics (CFD) simulations requiring weeks per design iteration, or simplified models that sacrifice production-grade accuracy. While machine learning offers transformative potential, existing datasets exhibit fundamental limitations -- inadequate mesh resolution, missing vehicle components, and validation errors exceeding 5% -- preventing deployment in industrial workflows. We present DrivAerStar, comprising 12,000 industrial-grade automotive CFD simulations generated using STAR-CCM+${}^{\textregistered}$ software. The dataset systematically explores three vehicle configurations through 20 Computer Aided Design (CAD) parameters via Free Form Deformation (FFD) algorithms, including complete engine compartments and cooling systems with realistic internal airflow. DrivAerStar achieves wind tunnel validation accuracy below 1.04% -- a five-fold improvement over existing datasets -- through refined mesh strategies with strict wall $y^+$ control. Benchmarks demonstrate that models trained on this data achieve production-ready accuracy while reducing computational costs from weeks to minutes. This represents the first dataset bridging academic machine learning research and industrial CFD practice, establishing a new standard for data-driven aerodynamic optimization in automotive development. Beyond automotive applications, DrivAerStar demonstrates a paradigm for integrating high-fidelity physics simulations with Artificial Intelligence (AI) across engineering disciplines where computational constraints currently limit innovation.

Computer Vision · Vision Models & Multimodal

Zizhao Chen, Yoav Artzi

We propose KnotGym, an interactive environment for complex, spatial reasoning and manipulation. KnotGym includes goal-oriented rope manipulation tasks with varying levels of complexity, all requiring acting from pure image observations.Tasks are defined along a clear and quantifiable axis of complexity based on the number of knot crossings, creating a natural generalization test.KnotGym has a simple observation space, allowing for scalable development, yet it highlights core challenges in integrating acute perception, spatial reasoning, and grounded manipulation.We evaluate methods of different classes, including model-based RL, model-predictive control, and chain-of-thought reasoning, and illustrate the challenges KnotGym presents.

Social Aspects · Privacy-preserving Statistics and Machine Learning

Wanying Dai, Beibei Li, Naipeng Dong, Guangdong Bai, Jin Song Dong

Face recognition is essential for identity authentication, but the rich visual clues in facial images pose significant privacy risks, highlighting the critical importance of privacy-preserving solutions. For instance, numerous studies have shown that generative models are capable of effectively performing reconstruction attacks that result in the restoration of original visual clues. To mitigate this threat, we introduce FracFace, a fractal-based privacy-preserving face recognition framework. This approach effectively weakens the visual clues that can be exploited by reconstruction attacks by disrupting the spatial structure in frequency domain features, while retaining the vital visual clues required for identity recognition. To achieve this, we craft a Frequency Channels Refining module that reduces sparsity in the frequency domain. It suppresses visual clues that could be exploited by reconstruction attacks, while preserving features indispensable for recognition, thus making these attacks more challenging. More significantly, we design a Frequency Fractal Mapping module that obfuscates deep representations by remapping refined frequency channels into a fractal-based privacy structure. By leveraging the self-similarity of fractals, this module preserves identity relevant features while enhancing defense capabilities, thereby improving the overall robustness of the protection scheme. Experiments conducted on multiple public face recognition benchmarks demonstrate that the proposed FracFace significantly reduces the visual recoverability of facial features, while maintaining high recognition accuracy, as well as the superiorities over state-of-the-art privacy protection approaches.

Social Aspects · Trustworthy Machine Learning

Chongyu Fan, Jiancheng Liu, Licong Lin, Jinghan Jia, Ruiqi Zhang, Song Mei, Sijia Liu

This work studies the problem of large language model (LLM) unlearning, aiming to remove unwanted data influences (e.g., copyrighted or harmful content) while preserving model utility. Despite the increasing demand for unlearning, a technically-grounded optimization framework is lacking. Gradient ascent (GA)-type methods, though widely used, are suboptimal as they reverse the learning process without controlling optimization divergence (i.e., deviation from the pre-trained state), leading to risks of model collapse. Negative preference optimization (NPO) has been proposed to address this issue and is considered one of the state-of-the-art LLM unlearning approaches. In this work, we revisit NPO and identify another critical issue: reference model bias. This bias arises from using the reference model (i.e., the model prior to unlearning) to assess unlearning success, which can lead to a misleading impression of the true data-wise unlearning effectiveness. Specifically, it could cause (a) uneven allocation of optimization power across forget data with varying difficulty levels, and (b) ineffective gradient weight smoothing during the early stages of unlearning optimization. To overcome these challenges, we propose a simple yet effective unlearning optimization framework, called SimNPO, showing that simplicity—removing the reliance on a reference model (through the lens of simple preference optimization)—benefits unlearning. We provide deeper insights into SimNPO's advantages, including an analysis based on mixtures of Markov chains. Extensive experiments further validate its efficacy on benchmarks like TOFU, MUSE, and WMDP.

Social Aspects · Privacy-preserving Statistics and Machine Learning

Guancheng Wan, Xu Cheng, Run Liu, Wenke Huang, Zitong Shi, Pinyi Jin, Guibin Zhang, Bo Du, Mang Ye

Federated Graph Learning (FGL) has been shown to be particularly effective in enabling collaborative training of Graph Neural Networks (GNNs) in decentralized settings. Model-heterogeneous FGL further enhances practical applicability by accommodating client preferences for diverse model architectures. However, existing model-heterogeneous approaches primarily target Euclidean data and fail to account for a crucial aspect of graph-structured data: topological relationships. To address this limitation, we propose **TRUST**, a novel knowledge distillation-based **model-heterogeneous FGL** framework. Specifically, we propose Progressive Curriculum Node Scheduler to progressively introduce challenging nodes based on learning difficulty. In Adaptive Curriculum Distillation Modulator, we propose an adaptive temperature modulator that dynamically adjusts knowledge distillation temperature to accommodate varying client capabilities and graph complexity. Moreover, we leverage Wasserstein‑Driven Affinity Distillation to enable models to capture cross-class structural relationships through optimal transport. Extensive experiments on multiple graph benchmarks and model-heterogeneous settings show that **TRUST** outperforms existing methods, achieving an average 3.6\% $\uparrow$ performance gain, particularly under moderate heterogeneity conditions. The code is available for anonymous access at https://anonymous.4open.science/r/TRUST-NeurIPS2025.

Deep Learning · Everything Else

Qinting Jiang, Chuyang Ye, Dongyan Wei, Bingli Wang, Yuan Xue, Jingyan Jiang, Zhi Wang

Despite progress, deep neural networks still suffer performance declines under distribution shifts between training and test domains, leading to a substantial decrease in Quality of Experience (QoE) for applications. Existing test-time adaptation (TTA) methods are challenged by dynamic, multiple test distributions within batches. We observe that feature distributions across different domains inherently cluster into distinct groups with varying means and variances. This divergence reveals a critical limitation of previous global normalization strategies in TTA, which inevitably distort the original data characteristics. Based on this insight, we propose Feature-based Instance Neighbor Discovery (FIND), which comprises three key components: Layer-Wise Feature Disentanglement (LFD), Feature-Aware Batch Normalization (FABN) and Selective FABN (S-FABN). LFD stably captures features with similar distributions at each layer by constructing graph structures; while FABN optimally combines source statistics with test-time distribution-specific statistics for robust feature representation. Finally, S-FABN determines which layers require feature partitioning and which can remain unified, thus enhancing the efficiency of inference. Extensive experiments demonstrate that FIND significantly outperforms existing methods, achieving up to approximately 30\% accuracy improvement in dynamic scenarios while maintaining computational efficiency. The source code is available at https://github.com/Peanut-255/FIND.

Deep Learning · Algorithms

Kaichen Zhang, Yuzhong Hong, Junwei Bao, Hongfei Jiang, Yang Song, Hong Dingqian, Hui Xiong

Post-training plays a crucial role in refining and aligning large language models to meet specific tasks and human preferences. While recent advancements in post-training techniques, such as Group Relative Policy Optimization (GRPO), leverage increased sampling with relative reward scoring to achieve superior performance, these methods often suffer from training instability that limits their practical adoption. As a next step, we present Group Variance Policy Optimization (GVPO). GVPO incorporates the analytical solution to KL-constrained reward maximization directly into its gradient weights, ensuring alignment with the optimal policy. The method provides intuitive physical interpretations: its gradient mirrors the mean squared error between the central distance of implicit rewards and that of actual rewards. GVPO offers two key advantages: (1) it guarantees a unique optimal solution, exactly the KL-constrained reward maximization objective, (2) it supports flexible sampling distributions that avoids on-policy and importance sampling limitations. By unifying theoretical guarantees with practical adaptability, GVPO establishes a new paradigm for reliable and versatile LLM post-training.

Applications · Robotics

Jared Perlo, Alexander Robey, Fazl Barez, Jakob Mökander

The field of embodied AI (EAI) is rapidly advancing. Unlike virtual AI, EAI systems can exist in, learn from, reason about, and act in the physical world. With recent advances in AI and hardware research and design, EAI systems are becoming increasingly capable across an expanding set of operational domains. While EAI systems can offer many benefits, they also pose significant short- and long-term risks, including physical harm, surveillance, and societal disruption. These risks require urgent attention from policymakers, as existing policies for industrial robots and autonomous vehicles are insufficient to manage the full range of concerns EAI systems present. To address this issue, this paper makes three contributions. First, we provide a taxonomy of the physical, informational, economic, and social risks EAI systems pose. Second, we analyze policies in the US, UK, and EU to assess how existing frameworks address these risks and to identify critical gaps. We conclude by offering policy recommendations for the safe and beneficial deployment of EAI systems, such as mandatory testing and certification schemes, clarified liability frameworks, and strategies to manage EAI’s potentially transformative economic and societal impacts.

Computer Vision · Classification and Understanding

Thomas Eiter, Katsumi Inoue, Nelson Higuera, Sota Moriyama

Integrating logical constraints into object detection models for autonomous driving (AD) is a promising way to enhance their compliance with rules and thereby increase the safety of the system. T-norms have been utilized to calculate the constrained loss, i.e., the violations of logical constraints as losses. While prior works have statically selected a few t-norms, we conduct an extensive experimental study to identify the most effective choices, as suboptimal t-norms can lead to undesired model behavior. To this end, we present MOD-ECL, a neurosymbolic framework that implements a wide range of t-norms and applies them in an adaptive manner. It includes an algorithm that selects well-performing t-norms during training and a scheduler that regulates the impact of the constrained loss. We evaluate its effectiveness on the ROAD-R and ROAD-Waymo-R datasets for object detection in AD, using attached common-sense constraints. Our results show that careful selection of parameters is crucial for effective constrained loss behavior. Moreover, our framework not only reduces constraint violations but also, in some cases, improves detection performance. Additionally, our methods offer fine-grained control over the trade-off between accuracy and constraint violation.

Computer Vision · Segmentation

Enzhi Zhang, Peng Chen, Rui Zhong, Du Wu, Jun Igarashi, Isaac Lyngaas, Xiao Wang, Masaharu Munetomo, Mohamed Wahib

This paper presents a novel approach to addressing the long-sequence problem in high-resolution medical images for Vision Transformers (ViTs). Using smaller patches as tokens can enhance ViT performance, but quadratically increases computation and memory requirements. Therefore, the common practice for applying ViTs to high-resolution images is either to: (a) employ complex sub-quadratic attention schemes or (b) use large to medium-sized patches and rely on additional mechanisms within the model to capture the spatial hierarchy of details. We propose Symmetrical Hierarchical Forest (SHF), a lightweight approach that adaptively patches the input image to increase token information density and encode hierarchical spatial structures into the input embedding. We then apply a reverse depatching scheme to the output embeddings of the transformer encoder, eliminating the need for convolution-based decoders. Unlike previous methods that modify attention mechanisms \wahib{or use a complex hierarchy of interacting models}, SHF can be retrofitted to any ViT model to allow it to learn the hierarchical structure of details in high-resolution images without requiring architectural changes. Experimental results demonstrate significant gains in computational efficiency and performance: on the PAIP WSI dataset, we achieved a 3$\sim$32$\times$ speedup or a 2.95\% to 7.03\% increase in accuracy (measured by Dice score) at a $64K^2$ resolution with the same computational budget, compared to state-of-the-art production models. On the 3D medical datasets BTCV and KiTS, training was 6$\times$ faster, with accuracy gains of 6.93\% and 5.9\%, respectively, compared to models without SHF.

Optimization · Large Scale, Parallel and Distributed

Hancheng Ye, Zhengqi Gao, Mingyuan Ma, Qinsi Wang, Yuzhe Fu, Ming-Yu Chung, Yueqian Lin, Zhijian Liu, Jianyi Zhang, Danyang Zhuo 等

Multi-agent large language model (LLM) systems are increasingly adopted for complex language processing tasks that require communication and coordination among agents. However, these systems often suffer substantial overhead from repeated reprocessing of overlapping contexts across agents. In typical pipelines, once an agent receives a message from its predecessor, the full context-including prior turns-must be reprocessed from scratch, leading to inefficient processing. While key-value (KV) caching is an effective solution for avoiding redundant computation in single-agent settings where prefixes remain unchanged, it cannot be directly reused in multi-agent scenarios due to diverging prefixes introduced by agent-specific context extensions. We identify that the core challenge lies in the offset variance of KV-caches across agents. To address this, we propose **KVCOMM**, a training-free framework that enables efficient prefilling in multi-agent inference by reusing KV-caches and aligning cache offsets of overlapping contexts under diverse prefix contexts. KVCOMM estimates and adjusts KV-caches for shared content by referencing a pool of cached examples—termed *anchors*—that store observed cache deviations under varying prefixes. The anchor pool is maintained and updated online, allowing dynamic adaptation to distinct user requests and context structures. KVCOMM achieves over 70% reuse rate across diverse multi- agent workloads, including retrieval-augmented generation, math reasoning, and collaborative coding tasks, all without quality degradation. Particularly, when each fully-connected agent receives 1K input tokens with 512 prefix tokens and 512 output tokens under a five-agent setting, KVCOMM achieves up to 7.8× speedup compared to the standard prefill pipeline, reducing TTFT from ∼430ms to ∼55ms. Code is available at https://github.com/FastMAS/KVCOMM.

Computer Vision · Classification and Understanding

Yujin Jeong, Arnas Uselis, Seong Joon Oh, Anna Rohrbach

Understanding visual scenes is fundamental to human intelligence. While discriminative models have significantly advanced computer vision, they often struggle with compositional understanding. In contrast, recent generative text-to-image diffusion models excel at synthesizing complex scenes, suggesting inherent compositional capabilities.Building on this, zero-shot diffusion classifiers have been proposed to repurpose diffusion models for discriminative tasks. While prior work offered promising results in discriminative compositional scenarios, these results remain preliminary due to a small number of benchmarks and a relatively shallow analysis of conditions under which the models succeed. To address this, we present a comprehensive study of the discriminative capabilities of diffusion classifiers on a wide range of compositional tasks. Specifically, our study covers three diffusion models (SD 1.5, 2.0, and, for the first time, 3-m) spanning 10 datasets and over 30 tasks. Further, we shed light on the role that target dataset domains play in respective performance; to isolate the domain effects, we introduce a new diagnostic benchmark \textsc{Self-Bench} comprised of images created by diffusion models themselves. Finally, we explore the importance of timestep weighting and uncover a relationship between domain gap and timestep sensitivity, particularly for SD3-m.To sum up, diffusion classifiers understand compositionality, but conditions apply! Code and dataset are available at https://github.com/eugene6923/Diffusion-Classifiers-Compositionality

Probabilistic Methods · Gaussian Processes

Pratik Rathore, Zachary Frangella, Sachin Garg, Shaghayegh Fazliani, Michal Derezinski, Madeleine Udell

Gaussian processes (GPs) play an essential role in biostatistics, scientific machine learning, and Bayesian optimization for their ability to provide probabilistic predictions and model uncertainty. However, GP inference struggles to scale to large datasets (which are common in modern applications), since it requires the solution of a linear system whose size scales quadratically with the number of samples in the dataset. We propose an approximate, distributed, accelerated sketch-and-project algorithm ($\texttt{ADASAP}$) for solving these linear systems, which improves scalability. We use the theory of determinantal point processes to show that the posterior mean induced by sketch-and-project rapidly converges to the true posterior mean. In particular, this yields the first efficient, condition number-free algorithm for estimating the posterior mean along the top spectral basis functions, showing that our approach is principled for GP inference. $\texttt{ADASAP}$ outperforms state-of-the-art solvers based on conjugate gradient and coordinate descent across several benchmark datasets and a large-scale Bayesian optimization task. Moreover, $\texttt{ADASAP}$ scales to a dataset with $> 3 \cdot 10^8$ samples, a feat which has not been accomplished in the literature.

Deep Learning · Algorithms

Panagiotis Giannoulis, Yorgos Pantis, Christos Tzamos

Despite their proficiency in various language tasks, Large Language Models (LLMs) struggle with combinatorial problems like Satisfiability, Traveling Salesman Problem, or even basic arithmetic. We address this gap through a novel approach for solving problems in the class NP. We focus on the paradigmatic task of Sudoku and achieve state-of-the-art accuracy (99\%) compared to prior neuro-symbolic approaches. Unlike prior work that used custom architectures, our method employs a vanilla decoder-only Transformer (GPT-2) without external tools or function calling. Our method integrates imitation learning of simple Sudoku rules with an explicit Depth-First Search (DFS) exploration strategy involving informed guessing and backtracking. Moving beyond imitation learning, we seek to minimize the number of guesses until reaching a solution. We provide a rigorous analysis of this setup by formalizing its connection to a contextual variant of $\textit{Min-Sum Set Cover}$, a well-studied problem in algorithms and stochastic optimization.