arXiv subjects

Yipeng Zhang

Publications and source records attributed to Yipeng Zhang.

At least 19 recordsLinked to original sources

arXiv2026

The Optimal Scaling Parameter in Metric Cotype for Alexandrov Spaces of Nonnegative Curvature

We determine the optimal order of the scaling parameter in the metric cotype $2$ inequality for complete Alexandrov spaces of nonnegative curvature. It grows linearly with the dimension $n$ of the discrete torus. This answers Question 17 of Eskenazis, Mendel, and Naor. In their sign-vector normalization on $\bigl(\mathbb{Z}/(2m\mathbb{Z})\bigr)^n$, the square of the optimal metric cotype $2$ constant for this class lies between $\max\{1,n/m\}$ and $e\lceil n/m\rceil$ for every dyadic $m\geq 2$. The upper bound follows from a dyadic Bernoulli-thinning argument using only the Lang--Schroeder--Sturm inequality. Snowflake universality of the fixed Wasserstein space $\mathcal{P}_2(\mathbb{R}^3)$ gives the $n/m$ lower bound.

math.FA↗

arXiv2026

Quadratic Convexification of a Square Truncated by a Hyperbola

We study the quadratic convexification of the compact nonconvex set ${G} := \left[1/2,2\right]^2 \cap \{(x_1,x_2)\in\mathbb R^2\mid x_1x_2\leq1\}, $ which is a square truncated by a hyperbolic arc. Although the ordinary convex hull of ${G}$ is a triangle, its lifted convex hull in the complete quadratic space retains the nontrivial geometry of the curved boundary. We characterize all extreme rays of the cone of quadratic polynomials nonnegative on $G$, including parameterized families of bounded tangent and bitangent rays. Using this classification and conic duality, we derive an exact finite semidefinite representation of the lifted convex hull of ${G}$. The analysis follows the general framework of our earlier work on an unbounded product-constrained region, but the bounded geometry creates new boundary-contact patterns and leads to a different finite organization of the nonnegative-quadratic cone.

math.OC↗

arXiv2026

Nonnegative Quadratics over a Quadrant with a Bilinear Constraint

We study quadratic polynomials that are nonnegative on the non-compact set \[ F:=\{(x_1,x_2)\in\mathbb R^2:\ x_1\ge 0,\ x_2\ge 0,\ x_1x_2\le 1\}. \] All extreme rays of the cone of nonnegative quadratic polynomials are characterized on this set. The characterization allows us to study parameterized valid inequalities for quadratic convexifications involving $F$, which yields a semidefinite representation of its lifted convex hull in the quadratic space. Our analysis on extreme ray characterization separates the positive-semidefinite (PSD) and non-PSD branches, reduces the latter to boundary nonnegativity, and classifies the boundary contacts of the relevant extreme rays. By reparameterization, we also find a non-trivial and non-permutation-symmetric six-dimensional linear section of the cone of nonnegative homogeneous ternary octics that are sum-of-squares. We also show that our lifted convex hull result yields a degree-bounded preordering certificate of nonnegative quadratics on $F$ and a degree-bounded certificate for a family of nonnegative quartics on the half-strip.

math.OC↗

arXiv2026

Communicability-Inspired Positional Encoding (CIPE)

Positional encodings (PEs) are essential for Transformers. Yet designing effective PEs for non-Euclidean graphs remains challenging. Such encodings should ideally induce an Attention-Compatible Geometry for self-attention: not merely describing graph structure, but defining a geometry whose inner products reflect meaningful structural relatedness. To realize this geometry, we propose Communicability-Inspired Positional Encoding (CIPE), built from communicability, a measure between pairs of nodes that aggregates contributions from paths of all lengths. By construction, CIPE inner products recover communicability, converting global multi-path connectivity into an attention-ready similarity geometry. For practical Transformer training, we introduce dimensionality alignment, mapping graph-size-dependent CIPE representations to prescribed dimensions while faithfully preserving the induced geometry. Empirically, CIPE improves structure-agnostic Transformers by 35.5% on average across seven benchmarks, outperforming representative PEs; it also consistently improves structure-biased graph Transformers, where competing PEs often yield only marginal benefits. These results position CIPE as a principled framework for attention-compatible graph positional encodings.

cs.LG↗

arXiv2026

RetroAgent: From Solving to Evolving via Retrospective Dual Intrinsic Feedback

Standard reinforcement learning (RL) for large language model (LLM) agents primarily optimizes extrinsic task rewards, often favoring isolated task completion over continual adaptation. This paradigm can cause premature convergence to suboptimal policies and leaves useful experience only implicitly encoded in model parameters, limiting its retrieval and reuse for future decisions. We introduce RetroAgent, an online RL framework that trains agents to master interactive environments not merely by solving tasks, but by evolving across episodes. Inspired by human retrospective self-improvement, RetroAgent augments extrinsic rewards with hindsight-generated dual intrinsic feedback: (1) Intrinsic Numerical Feedback, which rewards beneficial exploration by measuring incremental subtask progress relative to prior attempts; and (2) Intrinsic Language Feedback which distills successes and failures into reusable textual lessons for explicit experience reuse. To leverage these lessons effectively, we propose Similarity & Utility-Aware Upper Confidence Bound (SimUtil-UCB), a retrieval strategy that balances semantic relevance, historical utility, and exploration. Across four challenging agentic benchmarks, RetroAgent achieves new state-of-the-art performance, outperforming GRPO by +18.3% on ALFWorld, +15.4% on WebShop, +27.1% on Sokoban, and +8.9% on MineSweeper, while demonstrating strong test-time adaptation and out-of-distribution generalization.

cs.AI↗

arXiv2026

A Note on the Gradient-Evaluation Sequence in Accelerated Gradient Methods

Nesterov's accelerated gradient descent method (AGD) is a seminal deterministic first-order method known to achieve the optimal order of iteration complexity for solving convex smooth optimization problems. Two distinct sequences of iterates are included in the description of AGD: gradient evaluations are performed at one sequence, while approximate solutions are selected from the other. The iteration complexity on minimizing objective function value has been well-studied in the literature, but such analysis is almost always performed only at the approximate solution sequence. To the best of our knowledge, for projection-based AGD that solves problems with feasible sets, it is still an open research question whether the gradient evaluation sequence (when treated as approximate solutions) could also achieve the same optimal order of iteration complexity. It is also unknown whether such results still hold in the non-Euclidean setting. Motivated by computer-aided algorithm analysis, we provide positive results that answer the open problems affirmatively. Specifically, for (possibly constrained) problem $f^*:=\min_{x\in X}f(x)$ where $f$ is convex and $L$-smooth and $X$ is closed, convex and projection friendly, we prove that the gradient-evaluation sequence $\{\underline{x}_k\}$ in AGD satisfies that $f(\underline{x}_k) - f^*\le \mathcal O(L/k^2)$.

math.OC↗

arXiv2026

Omni-iEEG: A Large-Scale, Comprehensive iEEG Dataset and Benchmark for Epilepsy Research

Epilepsy affects over 50 million people worldwide, and one-third of patients suffer drug-resistant seizures where surgery offers the best chance of seizure freedom. Accurate localization of the epileptogenic zone (EZ) relies on intracranial EEG (iEEG). Clinical workflows, however, remain constrained by labor-intensive manual review. At the same time, existing data-driven approaches are typically developed on single-center datasets that are inconsistent in format and metadata, lack standardized benchmarks, and rarely release pathological event annotations, creating barriers to reproducibility, cross-center validation, and clinical relevance. With extensive efforts to reconcile heterogeneous iEEG formats, metadata, and recordings across publicly available sources, we present $\textbf{Omni-iEEG}$, a large-scale, pre-surgical iEEG resource comprising $\textbf{302 patients}$ and $\textbf{178 hours}$ of high-resolution recordings. The dataset includes harmonized clinical metadata such as seizure onset zones, resections, and surgical outcomes, all validated by board-certified epileptologists. In addition, Omni-iEEG provides over 36K expert-validated annotations of pathological events, enabling robust biomarker studies. Omni-iEEG serves as a bridge between machine learning and epilepsy research. It defines clinically meaningful tasks with unified evaluation metrics grounded in clinical priors, enabling systematic evaluation of models in clinically relevant settings. Beyond benchmarking, we demonstrate the potential of end-to-end modeling on long iEEG segments and highlight the transferability of representations pretrained on non-neurophysiological domains. Together, these contributions establish Omni-iEEG as a foundation for reproducible, generalizable, and clinically translatable epilepsy research. The project page with dataset and code links is available at omni-ieeg.github.io/omni-ieeg.

cs.LG↗

arXiv2026

Self-Supervised Learning from Structural Invariance

Joint-embedding self-supervised learning (SSL), the key paradigm for unsupervised representation learning from visual data, learns from invariances between semantically-related data pairs. We study the one-to-many mapping problem in SSL, where each datum may be mapped to multiple valid targets. This arises when data pairs come from naturally occurring generative processes, e.g., successive video frames. We show that existing methods struggle to flexibly capture this conditional uncertainty. As a remedy, we introduce a latent variable to account for this uncertainty and derive a variational lower bound on the mutual information between paired embeddings. Our derivation yields a simple regularization term for standard SSL objectives. The resulting method, which we call AdaSSL, applies to both contrastive and distillation-based SSL objectives, and we empirically show its versatility in causal representation learning, fine-grained image understanding, and world modeling on videos.

cs.LG↗

arXiv2025

Scalpel-SAM: A Semi-Supervised Paradigm for Adapting SAM to Infrared Small Object Detection

Infrared small object detection urgently requires semi-supervised paradigms due to the high cost of annotation. However, existing methods like SAM face significant challenges of domain gaps, inability of encoding physical priors, and inherent architectural complexity. To address this, we designed a Hierarchical MoE Adapter consisting of four white-box neural operators. Building upon this core component, we propose a two-stage paradigm for knowledge distillation and transfer: (1) Prior-Guided Knowledge Distillation, where we use our MoE adapter and 10% of available fully supervised data to distill SAM into an expert teacher (Scalpel-SAM); and (2) Deployment-Oriented Knowledge Transfer, where we use Scalpel-SAM to generate pseudo labels for training lightweight and efficient downstream models. Experiments demonstrate that with minimal annotations, our paradigm enables downstream models to achieve performance comparable to, or even surpassing, their fully supervised counterparts. To our knowledge, this is the first semi-supervised paradigm that systematically addresses the data scarcity issue in IR-SOT using SAM as the teacher model.

cs.CV↗

arXiv2025

Hybrid Quantum-Classical Neural Networks for Few-Shot Credit Risk Assessment

Quantum Machine Learning (QML) offers a new paradigm for addressing complex financial problems intractable for classical methods. This work specifically tackles the challenge of few-shot credit risk assessment, a critical issue in inclusive finance where data scarcity and imbalance limit the effectiveness of conventional models. To address this, we design and implement a novel hybrid quantum-classical workflow. The methodology first employs an ensemble of classical machine learning models (Logistic Regression, Random Forest, XGBoost) for intelligent feature engineering and dimensionality reduction. Subsequently, a Quantum Neural Network (QNN), trained via the parameter-shift rule, serves as the core classifier. This framework was evaluated through numerical simulations and deployed on the Quafu Quantum Cloud Platform's ScQ-P21 superconducting processor. On a real-world credit dataset of 279 samples, our QNN achieved a robust average AUC of 0.852 +/- 0.027 in simulations and yielded an impressive AUC of 0.88 in the hardware experiment. This performance surpasses a suite of classical benchmarks, with a particularly strong result on the recall metric. This study provides a pragmatic blueprint for applying quantum computing to data-constrained financial scenarios in the NISQ era and offers valuable empirical evidence supporting its potential in high-stakes applications like inclusive finance.

cs.LG↗

arXiv2025

LESER: Learning to Expand via Search Engine-feedback Reinforcement in e-Commerce

User queries in e-commerce search are often vague, short, and underspecified, making it difficult for retrieval systems to match them accurately against structured product catalogs. This challenge is amplified by the one-to-many nature of user intent, where a single query can imply diverse and competing needs. Existing methods, including neural query expansion and prompting-based LLM approaches, fall short in real-world settings: they struggle to capture nuanced user intent, often generate outputs that violate platform constraints, and rely on workflows that are difficult to scale in production. We propose Learning to Expand via Search Engine-feedback Reinforcement (LESER), a novel framework that fine-tunes a context-aware LLM using real-time search engine feedback as supervision. LESER formulates query expansion as a retrieval optimization task and leverages Group Relative Policy Optimization to learn directly from relevance and coverage metrics. LESER is trained to reason over search results and produce high quality query expansions that align with platform rules and retrieval objectives. We evaluate LESER on large-scale, real-world e-commerce datasets, demonstrating substantial improvements in both offline and online settings. Our results show that LESER not only enhances semantic coverage and retrieval relevance but also delivers measurable gains in user engagement, making it a practical and scalable solution for modern search systems.

cs.IR↗

arXiv2025

Text to Query Plans for Question Answering on Large Tables

Efficient querying and analysis of large tabular datasets remain significant challenges, especially for users without expertise in programming languages like SQL. Text-to-SQL approaches have shown promising performance on benchmark data; however, they inherit SQL's drawbacks, including inefficiency with large datasets and limited support for complex data analyses beyond basic querying. We propose a novel framework that transforms natural language queries into query plans. Our solution is implemented outside traditional databases, allowing us to support classical SQL commands while avoiding SQL's inherent limitations. Additionally, we enable complex analytical functions, such as principal component analysis and anomaly detection, providing greater flexibility and extensibility than traditional SQL capabilities. We leverage LLMs to iteratively interpret queries and construct operation sequences, addressing computational complexity by incrementally building solutions. By executing operations directly on the data, we overcome context length limitations without requiring the entire dataset to be processed by the model. We validate our framework through experiments on both standard databases and large scientific tables, demonstrating its effectiveness in handling extensive datasets and performing sophisticated data analyses.

cs.DB↗

arXiv2025

Non-Equilibrium Criticality-Enhanced Quantum Sensing with Superconducting Qubits

Exploiting quantum features allows for estimating external parameters with precisions well beyond the capacity of classical sensors, a phenomenon known as quantum-enhanced precision. Quantum criticality has been identified as a resource for achieving such enhancements with respect to the probe size. However, they demand complex probe preparation and measurement and the achievable enhancement is ultimately restricted to narrow parameter regimes. On the other hand, non-equilibrium probes harness dynamics, enabling quantum-enhanced precision with respect to time over a wide range of parameters through simple probe initialization. Here, we unify these approaches through a Stark-Wannier localization platform, where competition between a linear gradient field and particle tunneling enables quantum-enhanced sensitivity across an extended parameter regime. The probe is implemented on a 9-qubit superconducting quantum device, in both single- and double-excitation subspaces, where we explore its performance in the extended phase, the critical point and the localized phase. Despite employing only computational-basis measurements we have been able to achieve near-Heisenberg-limited precision by combining outcomes at distinct evolution times. In addition, we demonstrate that the performance of the probe in the entire extended phase is significantly outperforming the performance in the localized regime. Our results highlight Stark-Wannier systems as versatile platforms for quantum sensing, where the combination of criticality and non-equilibrium dynamics enhances precision over a wide range of parameters without stringent measurement requirements.

quant-ph↗

arXiv2025

Bridging Modality Gaps in e-Commerce Products via Vision-Language Alignment

Item information, such as titles and attributes, is essential for effective user engagement in e-commerce. However, manual or semi-manual entry of structured item specifics often produces inconsistent quality, errors, and slow turnaround, especially for Customer-to-Customer sellers. Generating accurate descriptions directly from item images offers a promising alternative. Existing retrieval-based solutions address some of these issues but often miss fine-grained visual details and struggle with niche or specialized categories. We propose Optimized Preference-Based AI for Listings (OPAL), a framework for generating schema-compliant, high-quality item descriptions from images using a fine-tuned multimodal large language model (MLLM). OPAL addresses key challenges in multimodal e-commerce applications, including bridging modality gaps and capturing detailed contextual information. It introduces two data refinement methods: MLLM-Assisted Conformity Enhancement, which ensures alignment with structured schema requirements, and LLM-Assisted Contextual Understanding, which improves the capture of nuanced and fine-grained information from visual inputs. OPAL uses visual instruction tuning combined with direct preference optimization to fine-tune the MLLM, reducing hallucinations and improving robustness across different backbone architectures. We evaluate OPAL on real-world e-commerce datasets, showing that it consistently outperforms baseline methods in both description quality and schema completion rates. These results demonstrate that OPAL effectively bridges the gap between visual and textual modalities, delivering richer, more accurate, and more consistent item descriptions. This work advances automated listing optimization and supports scalable, high-quality content generation in e-commerce platforms.

cs.IR↗

arXiv2025

VisGuard: Securing Visualization Dissemination through Tamper-Resistant Data Retrieval

The dissemination of visualizations is primarily in the form of raster images, which often results in the loss of critical information such as source code, interactive features, and metadata. While previous methods have proposed embedding metadata into images to facilitate Visualization Image Data Retrieval (VIDR), most existing methods lack practicability since they are fragile to common image tampering during online distribution such as cropping and editing. To address this issue, we propose VisGuard, a tamper-resistant VIDR framework that reliably embeds metadata link into visualization images. The embedded data link remains recoverable even after substantial tampering upon images. We propose several techniques to enhance robustness, including repetitive data tiling, invertible information broadcasting, and an anchor-based scheme for crop localization. VisGuard enables various applications, including interactive chart reconstruction, tampering detection, and copyright protection. We conduct comprehensive experiments on VisGuard's superior performance in data retrieval accuracy, embedding capacity, and security against tampering and steganalysis, demonstrating VisGuard's competence in facilitating and safeguarding visualization dissemination and information conveyance.

cs.CV↗

arXiv2025

Self-Supervised Distillation of Legacy Rule-Based Methods for Enhanced EEG-Based Decision-Making

High-frequency oscillations (HFOs) in intracranial Electroencephalography (iEEG) are critical biomarkers for localizing the epileptogenic zone in epilepsy treatment. However, traditional rule-based detectors for HFOs suffer from unsatisfactory precision, producing false positives that require time-consuming manual review. Supervised machine learning approaches have been used to classify the detection results, yet they typically depend on labeled datasets, which are difficult to acquire due to the need for specialized expertise. Moreover, accurate labeling of HFOs is challenging due to low inter-rater reliability and inconsistent annotation practices across institutions. The lack of a clear consensus on what constitutes a pathological HFO further challenges supervised refinement approaches. To address this, we leverage the insight that legacy detectors reliably capture clinically relevant signals despite their relatively high false positive rates. We thus propose the Self-Supervised to Label Discovery (SS2LD) framework to refine the large set of candidate events generated by legacy detectors into a precise set of pathological HFOs. SS2LD employs a variational autoencoder (VAE) for morphological pre-training to learn meaningful latent representation of the detected events. These representations are clustered to derive weak supervision for pathological events. A classifier then uses this supervision to refine detection boundaries, trained on real and VAE-augmented data. Evaluated on large multi-institutional interictal iEEG datasets, SS2LD outperforms state-of-the-art methods. SS2LD offers a scalable, label-efficient, and clinically effective strategy to identify pathological HFOs using legacy detectors.

cs.CE↗

arXiv2025

Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback

Recent advances in reinforcement learning (RL) using numerical rewards have significantly enhanced the complex reasoning capabilities of large language models (LLMs). However, we identify three fundamental limitations of purely numerical feedback: performance plateaus, ineffective spontaneous self-reflection, and persistent failures. We show that plateaued RL models can successfully refine failed solutions when given natural language critiques. Motivated by this, we propose Critique-GRPO, an online RL framework that integrates both natural language and numerical feedback for policy optimization. This approach enables LLMs to learn simultaneously from initial responses and critique-guided refinements, effectively internalizing the exploration benefits of both stages. Extensive experiments show that Critique-GRPO outperforms all compared supervised and RL-based fine-tuning methods, achieving average Pass@1 improvements of approximately +15.0-21.6% on various Qwen models and +7.3% on Llama-3.2-3B-Instruct across eight challenging reasoning tasks. Notably, Critique-GRPO facilitates effective self-improvement through self-critiquing, achieving substantial gains over GRPO, e.g., a +16.7% Pass@1 improvement on AIME 2024. The code and models are released at: https://github.com/zhangxy-2019/critique-GRPO

cs.CL↗

arXiv2025

Rhomboid Tiling for Geometric Graph Deep Learning

Graph Neural Networks (GNNs) have proven effective for learning from graph-structured data through their neighborhood-based message passing framework. Many hierarchical graph clustering pooling methods modify this framework by introducing clustering-based strategies, enabling the construction of more expressive and powerful models. However, all of these message passing framework heavily rely on the connectivity structure of graphs, limiting their ability to capture the rich geometric features inherent in geometric graphs. To address this, we propose Rhomboid Tiling (RT) clustering, a novel clustering method based on the rhomboid tiling structure, which performs clustering by leveraging the complex geometric information of the data and effectively extracts its higher-order geometric structures. Moreover, we design RTPool, a hierarchical graph clustering pooling model based on RT clustering for graph classification tasks. The proposed model demonstrates superior performance, outperforming 21 state-of-the-art competitors on all the 7 benchmark datasets.

cs.LG↗