{"total":16,"items":[{"citing_arxiv_id":"2607.02508","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"From SRA to Self-Flow: Data Augmentation or Self-Supervision?","primary_cat":"cs.CV","submitted_at":"2026-07-02T17:59:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Attention Separation ablations show that gains from SRA to Self-Flow in diffusion transformers arise mainly from noise-dimension data augmentation rather than token-level self-supervision.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.31204","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"AC3S: Adaptive Conditioning for 3D-Aware Synthetic Data Generation","primary_cat":"cs.CV","submitted_at":"2026-06-30T06:33:55+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"AC3S adds a self-supervised visual prompt modulator to ControlNet diffusion and a multi-agent VLM prompt composer to generate photorealistic images with accurate 2D/3D annotations while avoiding over-conditioning.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.28419","ref_index":8,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"MedDiffuseMix: Preserving Diagnostic Evidence with Saliency-Aware Diffusion Medical Image Data Augmentation","primary_cat":"cs.CV","submitted_at":"2026-06-25T14:57:12+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Saliency-guided diffusion mixing that preserves Grad-CAM-highlighted diagnostic regions improves medical image classification accuracy and AUC across four public datasets.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.08802","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Active Flow Expansion for Out-of-Distribution Discovery: from Theory to Molecules","primary_cat":"cs.LG","submitted_at":"2026-06-07T19:43:22+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"ActFlow expands the generable set of pre-trained flow models for out-of-distribution molecular and sequence design via active synthetic data generation and verifier feedback, with new statistical guarantees.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.04306","ref_index":122,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Organizational Control Layer: Governance Infrastructure at the Execution Boundary of LLM Agent Systems","primary_cat":"cs.MA","submitted_at":"2026-06-03T00:25:56+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.01825","ref_index":143,"ref_count":2,"confidence":0.9,"is_internal_anchor":false,"paper_title":"ROGLE: Robust Global-Local Alignment with Automated Region Supervision for Text-Based Person Search","primary_cat":"cs.CV","submitted_at":"2026-06-01T07:41:44+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"ROGLE introduces automated pseudo region-sentence pairs via RSM and multi-granular learning to boost fine-grained alignment in text-based person search, plus the P-VLG benchmark with over 100k annotated regions.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.26353","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Personalized Generative Models for Contextual Debiasing","primary_cat":"cs.CV","submitted_at":"2026-05-25T21:58:15+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"DecoupleGen personalizes diffusion models to create images with uncommon contexts for debiasing object recognition, yielding consistent gains on scene classification tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.19289","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"What Makes Synthetic Data Effective in Image Segmentation","primary_cat":"cs.CV","submitted_at":"2026-05-19T03:07:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Dense scene composition and instance fidelity in synthetic diffusion images drive better segmentation performance; SENSE framework exploits this to improve models on Cityscapes, COCO, and ADE20K.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.11231","ref_index":3,"ref_count":2,"confidence":0.9,"is_internal_anchor":false,"paper_title":"LiBaGS: Lightweight Boundary Gap Synthesis for Targeted Synthetic Data Selection","primary_cat":"cs.LG","submitted_at":"2026-05-11T20:46:35+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"LiBaGS scores and selects synthetic data near decision boundaries using proximity, uncertainty, density, and validity, with boundary-gap allocation and marginal stopping to improve training accuracy.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.02583","ref_index":21,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Stylistic Attribute Control in Latent Diffusion Models","primary_cat":"cs.CV","submitted_at":"2026-05-04T13:34:14+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A technique for parametric stylistic control in latent diffusion models learns disentangled directions from synthetic datasets and applies them via guidance composition while preserving semantics.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"In contrast to deep learning, traditional image-based artistic ren- dering (IB-AR) methods [KCWI13] offer granular control through a series of engineered filters designed for specific artistic styles. Although highly controllable, these methods are constrained to specific, handcrafted styles like cartoon, oilpainting or watercolor effects [WOG06, SLKD16, BKTS06] or stroke-based [LLH ∗21, ZSQ∗21] or learnable filter-based [LRB ∗22, RBB∗24] de- and re- composition of images. Further, their post-hoc filtering is not in- formed of high level semantics and user intent that the text-to- image generative methods are given, such as prompt and spatial guidance, reducing their ability for content- and style aware ad- justments. We make use of IB-AR for synthetic data generation"},{"citing_arxiv_id":"2604.18076","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Class-specific diffusion models improve military object detection in a low-data domain","primary_cat":"cs.CV","submitted_at":"2026-04-20T10:46:41+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Class-specific diffusion models fine-tuned on 8-24 real images per class generate synthetic data that improves military vehicle detection by up to 8% mAP50 in low-data regimes, with further gains from ControlNet edge conditioning.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.12335","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"All in One: A Unified Synthetic Data Pipeline for Multimodal Video Understanding","primary_cat":"cs.CV","submitted_at":"2026-04-14T06:17:35+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"A unified synthetic data generation pipeline produces unlimited annotated multimodal video data across multiple tasks, enabling models trained mostly on synthetic data to generalize effectively to real-world video understanding benchmarks.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"The final output is a high-fidelity Generated video w/wo audio suitable for training MLLMs on complex temporal and cross-modal reasoning tasks. datasets [71, 83], but the resulting distributions often di- verge from real-world data. Recent advances in genera- tive models have narrowed this gap, enabling higher-fidelity synthetic data for both language [28, 39, 68, 96, 108] and vision tasks [2, 5, 8, 34, 37]. However, synthetic data can be noisy and inconsistently aligned, requiring noise-robust learning strategies [57, 58, 97], particularly in contrastive learning settings [46, 64]. Prior work on synthetic compo- sitional training highlights challenges in generating precise variations and maintaining cross-modal consistency. Our approach differs in three key aspects."},{"citing_arxiv_id":"2509.26158","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Towards Continual Expansion of Data Coverage: Automatic Text-guided Edge-case Synthesis","primary_cat":"cs.CV","submitted_at":"2025-09-30T12:11:25+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Automated LLM-based prompt engineering for text-to-image edge-case synthesis improves object detection robustness on the FishEye8K benchmark over naive augmentation and manual prompts.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2504.19455","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Masked Language Prompting for Generative Data Augmentation in Few-shot Fashion Style Recognition","primary_cat":"cs.CV","submitted_at":"2025-04-28T03:42:42+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Masked Language Prompting masks selected words in reference captions and leverages LLMs to produce diverse, semantically coherent completions for style-consistent generative image augmentation without fine-tuning.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2310.06114","ref_index":240,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Learning Interactive Real-World Simulators","primary_cat":"cs.AI","submitted_at":"2023-10-09T19:42:22+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"UniSim learns a universal real-world simulator from orchestrated diverse datasets, enabling zero-shot deployment of policies trained purely in simulation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2309.16797","ref_index":37,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution","primary_cat":"cs.CL","submitted_at":"2023-09-28T19:01:07+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Discuss the history, impact, and current status of [PROMPT1]; Therefore, the answer (arabic numerals) is 11 Discus Discuss the history, impact, and current status of [PROMPT1 Tell me about the history of [PROMPT1]; Discuss the impact of [ PROMPT1]; Give me the current status of [PROMPT1]; Tell me about the history of [PROMPT2]; Discuss the impact of [PROMPT2 Therefore, the answer (arabic numerals) is 11 Discus Context 2: Q. Paco ate 36 cookies. If he had 37 cookies initially How many cookies did Paco have left? A. 154 = 72 + 82. A. Paco ate 36 cookies. He had 37 cookies initially. This means he has 37 - 36 = 1 cookie left. So the answer is 1 Discuss the history, impact, and current status of [PROMPT1]; Therefore, the answer (arabic numerals) is 1 Discus Discuss the history, impact, and current status of [PROMPT1]; Therefore, the answer (arabic numerals) is 1 Discus 34 K APE I NSTRUCTION INDUCTION TASKS To demonstrate Promptbreeder's ability to evolve few-shot contexts as well as task-prompts we ran few-shot Promptbreeder on all 24 Instruction Induction datasets used in the APE e xperiments."}],"limit":50,"offset":0}