llms.txt
Compare with other sites' llms.txt files →# ArxivLens - Discover Breakthrough Research Instantly
> ArxivLens helps researchers stop wasting hours on research and find any paper in seconds. Search across major databases (arXiv, PubMed, bioRxiv, medRxiv, ACM), get AI insights instantly, and save 10+ hours per week on literature reviews.
ArxivLens transforms academic research discovery by providing unified search across multiple databases with AI-powered paper analysis.
## Core Features
- [Search Papers](https://arxivlens.com/search): Unified search across arXiv, PubMed, bioRxiv, and medRxiv with instant results
- [AI Paper Comparison](https://arxivlens.com/comparison): Compare multiple research papers side-by-side with AI analysis
- [Weekly Research Summaries](https://arxivlens.com/research/weekly-summaries): Curated weekly summaries
## Sitemap
- Sitemap index: https://arxivlens.com/sitemap.xml
## Most-Cited Papers in November 2026
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxivlens.com/PaperView/Details/very-deep-convolutional-networks-for-large-scale-image-recognition-5551-562188) (99981 citations): Researchers found that very deep convolutional networks achieve significant improvements in large-scale image recognition accuracy, achieving state-of-the-art results on multiple datasets. Their fi...
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://arxivlens.com/PaperView/Details/u-net-convolutional-networks-for-biomedical-image-segmentation-2320-633076) (76547 citations): A new neural network architecture called U-Net uses data augmentation to efficiently segment biomedical images from limited annotated samples, achieving better performance than previous methods.
- [Electric Field Effect in Atomically Thin Carbon Films](https://arxivlens.com/PaperView/Details/electric-field-effect-in-atomically-thin-carbon-films-6486-1145965) (56527 citations)
- [Going Deeper with Convolutions](https://arxivlens.com/PaperView/Details/going-deeper-with-convolutions-2504-565474) (43509 citations)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at
Scale](https://arxivlens.com/PaperView/Details/an-image-is-worth-16x16-words-transformers-for-image-recognition-at-scale-1530-1989580) (40650 citations): Researchers demonstrate that a pure transformer can perform well on image classification tasks, outperforming traditional convolutional networks with fewer computational resources. This breakthroug...
- [Convex optimization](https://arxivlens.com/PaperView/Details/convex-optimization-1647-1878147) (39856 citations): This textbook on convex optimization covers advanced topics such as conic optimization and robust optimization, presenting a unique perspective on the field.
- [SMOTE: Synthetic Minority Over-sampling Technique](https://arxivlens.com/PaperView/Details/smote-synthetic-minority-over-sampling-technique-8534-275655) (25432 citations): The SMOTE method improves classifier performance on imbalanced datasets by combining synthetic minority over-sampling with under-sampling of the majority class, outperforming other methods in terms...
- [Image-to-Image Translation with Conditional Adversarial Networks](https://arxivlens.com/PaperView/Details/image-to-image-translation-with-conditional-adversarial-networks-9127-800658) (19584 citations): Researchers developed a method to translate images using conditional adversarial networks, achieving success in tasks like image synthesis, object reconstruction, and colorization without requiring...
- [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxivlens.com/PaperView/Details/mobilenetv2-inverted-residuals-and-linear-bottlenecks-1770-2424501) (19281 citations): MobileNetV2 is a new mobile architecture that enhances performance in multiple tasks and model sizes with inverted residuals and linear bottlenecks. It also introduces efficient methods for object ...
- [YOLOv4: Optimal Speed and Accuracy of Object Detection](https://arxivlens.com/PaperView/Details/yolov4-optimal-speed-and-accuracy-of-object-detection-3981-2082122) (12216 citations)
- [Negative Refraction Makes a Perfect Lens](https://arxivlens.com/PaperView/Details/negative-refraction-makes-a-perfect-lens-3215-fa0764b1) (11065 citations): The paper proposes a 'super lens' using a negative refractive index material that can focus all parts of a 2D image, including non-radiative components, surpassing conventional lens limitations set...
- [Perceptual Losses for Real-Time Style Transfer and Super-Resolution](https://arxivlens.com/PaperView/Details/perceptual-losses-for-real-time-style-transfer-and-super-resolution-2686-725783) (10210 citations): A new approach to real-time style transfer and super-resolution image transformations combines the benefits of per-pixel and perceptual loss functions, achieving similar results but with significan...
- [Enriching Word Vectors with Subword Information](https://arxivlens.com/PaperView/Details/enriching-word-vectors-with-subword-information-1767-760364) (9931 citations): This paper introduces a new method for word representation, using character n-grams to enrich word vectors, overcoming the limitations of traditional models that treat each word as a unique vector....
- [Machine Learning in Automated Text Categorization](https://arxivlens.com/PaperView/Details/machine-learning-in-automated-text-categorization-1368-1102282) (8893 citations): This research paper explores machine learning techniques for automated text categorization, highlighting advances in document representation, classifier construction, and evaluation. The study aims...
- [A fast quantum mechanical algorithm for database search](https://arxivlens.com/PaperView/Details/a-fast-quantum-mechanical-algorithm-for-database-search-1284-853428) (8499 citations): A quantum algorithm can find a specific phone number in an N-name directory with only O(sqrt(N)) steps, significantly faster than classical algorithms.
- [How transferable are features in deep neural networks?](https://arxivlens.com/PaperView/Details/how-transferable-are-features-in-deep-neural-networks-5108-579971) (8309 citations): A study found that features learned in early layers of deep neural networks are generalizable across tasks, but those in later layers are less transferable due to specialization and optimization is...
- [Progressive Growing of GANs for Improved Quality, Stability, and
Variation](https://arxivlens.com/PaperView/Details/progressive-growing-of-gans-for-improved-quality-stability-and-variation-7070-2452234) (7356 citations): This paper introduces a new training method for Generative Adversarial Networks (GANs) that progressively grows the generator and discriminator, improving image quality, training stability, and var...
- [Deep Unsupervised Learning using Nonequilibrium Thermodynamics](https://arxivlens.com/PaperView/Details/deep-unsupervised-learning-using-nonequilibrium-thermodynamics-4757-614506) (6780 citations): This research introduces a novel deep unsupervised learning method inspired by non-equilibrium statistical physics. It systematically destroys data structure via an iterative forward diffusion proc...
- [Learning both Weights and Connections for Efficient Neural Networks](https://arxivlens.com/PaperView/Details/learning-both-weights-and-connections-for-efficient-neural-networks-3880-639277) (6628 citations): A new method reduces neural network parameters by an order of magnitude without affecting accuracy, making them more deployable on embedded systems. This is achieved through a three-step process: l...
- [An Alternative to Compactification](https://arxivlens.com/PaperView/Details/an-alternative-to-compactification-7479-968895) (5637 citations)
- [Evaluating Large Language Models Trained on Code](https://arxivlens.com/PaperView/Details/evaluating-large-language-models-trained-on-code-3494-1860478) (5393 citations): This paper presents Codex, a GPT model fine-tuned on GitHub code, which demonstrates proficient Python coding abilities. It solves 70.2% of problems on HumanEval, a new evaluation set, using repeat...
- [LINE: Large-scale Information Network Embedding](https://arxivlens.com/PaperView/Details/line-large-scale-information-network-embedding-2666-614499) (5318 citations): A new graph embedding method called LINE is proposed to efficiently embed large information networks into low-dimensional spaces, preserving local and global structures, and achieving impressive re...
- [EfficientDet: Scalable and Efficient Object Detection](https://arxivlens.com/PaperView/Details/efficientdet-scalable-and-efficient-object-detection-3510-2150156) (5023 citations): This paper introduces EfficientDet, a family of object detectors that achieve state-of-the-art efficiency while maintaining performance, with models 4-42x smaller in size and FLOPs compared to prio...
- [Spectral Networks and Locally Connected Networks on Graphs](https://arxivlens.com/PaperView/Details/spectral-networks-and-locally-connected-networks-on-graphs-2719-495763) (4856 citations)
- [HEALPix -- a Framework for High Resolution Discretization, and Fast
Analysis of Data Distributed on the Sphere](https://arxivlens.com/PaperView/Details/healpix-a-framework-for-high-resolution-discretization-and-fast-analysis-of-data-distributed-on-the-sphere-8314-1227978) (4592 citations): The HEALPix framework provides a versatile data structure for fast analysis of large astronomical datasets distributed on the sphere, addressing challenges for future missions like Planck and its p...
## Recent Highlights
- [Reversible Dynamics, Asymmetric Actions: Parking as a Trajectory-Level Affordance](https://arxivlens.com/PaperView/Details/reversible-dynamics-asymmetric-actions-parking-as-a-trajectory-level-affordance-ee643ebbdc9e7691) — published 2026-11-05
- [Time-Indexed Skill Success Models for Latency-Aware Skill Selection and Timing in Dynamic Manipulation](https://arxivlens.com/PaperView/Details/time-indexed-skill-success-models-for-latency-aware-skill-selection-and-timing-in-dynamic-manipulation-b8a1ea7d5f681832) — published 2026-11-05
- [Multiplicity via symmetry for the Minkowski mean curvature equation: Dirichlet and Neumann problems](https://arxivlens.com/PaperView/Details/multiplicity-via-symmetry-for-the-minkowski-mean-curvature-equation-dirichlet-and-neumann-problems-425e6a209e62a453) — published 2026-11-05
- [TCSExam: Evaluating Frontier LLMs on the Full Range of TCS Research](https://arxivlens.com/PaperView/Details/tcsexam-evaluating-frontier-llms-on-the-full-range-of-tcs-research-e122a2f22d39abc0) — published 2026-11-05
- [Bri3D: Bridging the Object-Scene Gap in Single-Image 3D Generation via Scene-Anchor](https://arxivlens.com/PaperView/Details/bri3d-bridging-the-object-scene-gap-in-single-image-3d-generation-via-scene-anchor-89bb869c46f2dd63) — published 2026-11-05
- [Monopoles, Scattering and Generalized Symmetries on a Quantum Computer](https://arxivlens.com/PaperView/Details/monopoles-scattering-and-generalized-symmetries-on-a-quantum-computer-0d446c723e6f4fd8) — published 2026-11-05
- [CoRD: Coupled Rollouts Fill the Blind Spot of On-Policy Distillation](https://arxivlens.com/PaperView/Details/cord-coupled-rollouts-fill-the-blind-spot-of-on-policy-distillation-7b7c116b0fe790c4) — published 2026-11-05
- [On the Spectral Dependence of Decentralized Stochastic Gradient Tracking](https://arxivlens.com/PaperView/Details/on-the-spectral-dependence-of-decentralized-stochastic-gradient-tracking-b7ae26a3f126cb77) — published 2026-11-05
- [Can $n_s$ hint at a parity-odd sector of gravity?](https://arxivlens.com/PaperView/Details/can-n-s-hint-at-a-parity-odd-sector-of-gravity-509e18126674dcdb) — published 2026-11-05
- [A Note on the Cut-Locus-Corrected Fréchet Mean Central Limit Theorem](https://arxivlens.com/PaperView/Details/a-note-on-the-cut-locus-corrected-frechet-mean-central-limit-theorem-cf7c05fd19797a1c) — published 2026-11-05