ciwGAN and fiwGAN models trained on isolated words spontaneously generate concatenated multi-word outputs and display early compositionality precursors.
hub Canonical reference
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
Canonical reference. 71% of citing Pith papers cite this work as background.
64
Pith papers citing it
Background
71% of classified citations
abstract
In recent years, supervised learning with convolutional networks (CNNs) has seen huge adoption in computer vision applications. Comparatively, unsupervised learning with CNNs has received less attention. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised learning. We introduce a class of CNNs called deep convolutional generative adversarial networks (DCGANs), that have certain architectural constraints, and demonstrate that they are a strong candidate for unsupervised learning. Training on various image datasets, we show convincing evidence that our deep convolutional adversarial pair learns a hierarchy of representations from object parts to scenes in both the generator and discriminator. Additionally, we use the learned features for novel tasks - demonstrating their applicability as general image representations.
hub tools
citation-role summary
background 5
method 2
citation-polarity summary
representative citing papers
Toy models demonstrate that polysemanticity arises when neural networks store more sparse features than neurons via superposition, producing a phase transition tied to polytope geometry and increased adversarial vulnerability.
GPT-f, a transformer-based prover for Metamath, generated new short proofs that were accepted into the main library—the first such contribution from a deep-learning system.
AGAN is the first neural architecture search method for GANs that discovers architectures outperforming state-of-the-art on CIFAR-10 unsupervised image generation and competitive on supervised tasks.
Real NVP uses affine coupling layers to create invertible transformations that support exact density estimation, sampling, and latent inference without approximations.
Synthetic minority augmentation improves threshold-integrated and optimized classification metrics only under model misspecification by correcting ranking errors, while providing no fundamental benefit beyond possible variance reduction under well-specified score models.
A generative-model-based test for equality of conditional distributions that uses cross-generation, an RKHS-indexed supremum statistic, and multiplier bootstrap, with claimed double robustness to generator errors.
A relative projection error metric in foundation-model embedding space predicts the downstream utility of synthetic positive samples for binary classifiers.
Prompts can be split into separate roles for sampling design and recovery modeling in generative compressed sensing, with stable recovery bounds for matched prompts and an explicit penalty for mismatch, validated on Stable Diffusion.
SurFITR is a new collection of 137k+ surveillance-style forged images that causes existing detectors to degrade while enabling substantial gains when used for training in both in-domain and cross-domain settings.
A pre-training diagnostic map based on spectral correlation resemblance to IQP circuits and excess structural complexity identifies suitable datasets like turbulence data for quantum generative models, yielding competitive low-resource performance.
ASTRA is a plug-and-play training-free method for precise multi-subject video editing that uses prompt-guided multimodal alignment and prior-based mask retargeting to avoid attention dilution and boundary issues.
Progressive growing stabilizes GAN training to produce high-resolution images of unprecedented quality and achieves a record unsupervised inception score of 8.80 on CIFAR10.
Mixed precision training uses FP16 for most computations, FP32 master weights for accumulation, and loss scaling to enable accurate training of large DNNs with halved memory usage.
AIS framework with time-varying piecewise parameters and hidden states enables analog hardware to perform generative modeling, achieving FID 27.6 on MNIST and 80.8 on Fashion-MNIST with 23uJ per image using 4-bit sparse architectures.
Integrates a diffusion prior into Richardson-Lucy deconvolution to reduce noise amplification and better preserve filamentous and punctate structures in low-photon fluorescence microscopy.
A fine-tuning-free framework combines pretrained Stable Diffusion with IP-Adapter plus three parameter-free modules to achieve improved lip synchronization and visual quality in talking face generation.
Sparse autoencoders scaled to 34 million features on Claude 3 Sonnet yield interpretable, steerable representations of concrete and abstract concepts that generalize across languages and modalities.
A systematic audit of 155 face swap apps finds 70% lack safeguards against generating non-consensual nude imagery.
GAI-GS integrates 3D Gaussian splatting and geometric algebra attention into an end-to-end model for wireless channel prediction that outperforms baselines on real indoor datasets.
VFMTok builds a generalist image tokenizer on frozen VFMs using adaptive quantization and semantic alignment, delivering gFID 1.36 for autoregressive and 1.25 for continuous generation on ImageNet with 3x faster convergence.
NeTMY neural fields with annealed encoding, multiscale optimization, and spectrum-fidelity losses achieve superior localization and distributional accuracy in NV-center inverse sensing by using a tensor power-summed dipolar operator that exposes and mitigates center-collapse failures.
W-Flow compresses a Wasserstein gradient flow defined via Sinkhorn divergence into a single-step neural generator, reporting 1.29 FID on ImageNet 256x256 with improved mode coverage.
Recursive generative retraining with heterogeneous rewards converges to a stable distribution satisfying a weighted Nash bargaining solution, preserving diversity under stated conditions.
citing papers explorer
-
Basic syntax from speech: Spontaneous concatenation in unsupervised deep neural networks
ciwGAN and fiwGAN models trained on isolated words spontaneously generate concatenated multi-word outputs and display early compositionality precursors.
-
Toy Models of Superposition
Toy models demonstrate that polysemanticity arises when neural networks store more sparse features than neurons via superposition, producing a phase transition tied to polytope geometry and increased adversarial vulnerability.
-
Generative Language Modeling for Automated Theorem Proving
GPT-f, a transformer-based prover for Metamath, generated new short proofs that were accepted into the main library—the first such contribution from a deep-learning system.
-
AGAN: Towards Automated Design of Generative Adversarial Networks
AGAN is the first neural architecture search method for GANs that discovers architectures outperforming state-of-the-art on CIFAR-10 unsupervised image generation and competitive on supervised tasks.
-
Density estimation using Real NVP
Real NVP uses affine coupling layers to create invertible transformations that support exact density estimation, sampling, and latent inference without approximations.
-
When Does Synthetic Data Augmentation Improve Score-Based Imbalanced Classification?
Synthetic minority augmentation improves threshold-integrated and optimized classification metrics only under model misspecification by correcting ranking errors, while providing no fundamental benefit beyond possible variance reduction under well-specified score models.
-
Testing Equality of Conditional Distributions via Generative Models
A generative-model-based test for equality of conditional distributions that uses cross-generation, an RKHS-indexed supremum statistic, and multiplier bootstrap, with claimed double robustness to generator errors.
-
Discriminative Span as a Predictor of Synthetic Data Utility via Classifier Reconstruction
A relative projection error metric in foundation-model embedding space predicts the downstream utility of synthetic positive samples for binary classifiers.
-
Active Learning for Conditional Generative Compressed Sensing
Prompts can be split into separate roles for sampling design and recovery modeling in generative compressed sensing, with stable recovery bounds for matched prompts and an explicit penalty for mismatch, validated on Stable Diffusion.
-
SurFITR: A Dataset for Surveillance Image Forgery Detection and Localisation
SurFITR is a new collection of 137k+ surveillance-style forged images that causes existing detectors to degrade while enabling substantial gains when used for training in both in-domain and cross-domain settings.
-
Toward Generative Quantum Utility via Correlation-Complexity Map
A pre-training diagnostic map based on spectral correlation resemblance to IQP circuits and excess structural complexity identifies suitable datasets like turbulence data for quantum generative models, yielding competitive low-resource performance.
-
ASTRA: Let Arbitrary Subjects Transform in Video Editing
ASTRA is a plug-and-play training-free method for precise multi-subject video editing that uses prompt-guided multimodal alignment and prior-based mask retargeting to avoid attention dilution and boundary issues.
-
Progressive Growing of GANs for Improved Quality, Stability, and Variation
Progressive growing stabilizes GAN training to produce high-resolution images of unprecedented quality and achieves a record unsupervised inception score of 8.80 on CIFAR10.
-
Mixed Precision Training
Mixed precision training uses FP16 for most computations, FP32 master weights for accumulation, and loss scaling to enable accurate training of large DNNs with halved memory usage.
-
Generative Models on Analog Hardware with Dynamics
AIS framework with time-varying piecewise parameters and hidden states enables analog hardware to perform generative modeling, achieving FID 27.6 on MNIST and 80.8 on Fashion-MNIST with 23uJ per image using 4-bit sparse architectures.
-
Improving Richardson--Lucy Deconvolution with Diffusion Priors for Fluorescence Microscopy
Integrates a diffusion prior into Richardson-Lucy deconvolution to reduce noise amplification and better preserve filamentous and punctate structures in low-photon fluorescence microscopy.
-
IP-Adapter Is All You Need: Towards Fine-Tuning-Free Diffusion-Based Talking Face Generation
A fine-tuning-free framework combines pretrained Stable Diffusion with IP-Adapter plus three parameter-free modules to achieve improved lip synchronization and visual quality in talking face generation.
-
Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet
Sparse autoencoders scaled to 34 million features on Claude 3 Sonnet yield interpretable, steerable representations of concrete and abstract concepts that generalize across languages and modalities.
-
Dual-Use AI Face Swap Apps Are Mostly Unsafe: A Systematic Safety Audit
A systematic audit of 155 face swap apps finds 70% lack safeguards against generating non-consensual nude imagery.
-
A Geometric Algebra-Informed 3DGS Framework for Wireless Channel Prediction
GAI-GS integrates 3D Gaussian splatting and geometric algebra attention into an end-to-end model for wireless channel prediction that outperforms baselines on real indoor datasets.
-
Vision Foundation Models as Generalist Tokenizers for Image Generation
VFMTok builds a generalist image tokenizer on frozen VFMs using adaptive quantization and semantic alignment, delivering gFID 1.36 for autoregressive and 1.25 for continuous generation on ImageNet with 3x faster convergence.
-
Neural Fields for NV-Center Inverse Sensing
NeTMY neural fields with annealed encoding, multiscale optimization, and spectrum-fidelity losses achieve superior localization and distributional accuracy in NV-center inverse sensing by using a tensor power-summed dipolar operator that exposes and mitigates center-collapse failures.
-
One-Step Generative Modeling via Wasserstein Gradient Flows
W-Flow compresses a Wasserstein gradient flow defined via Sinkhorn divergence into a single-step neural generator, reporting 1.29 FID on ImageNet 256x256 with improved mode coverage.
-
Curated Synthetic Data Doesn't Have to Collapse: A Theoretical Study of Generative Retraining with Pluralistic Preferences
Recursive generative retraining with heterogeneous rewards converges to a stable distribution satisfying a weighted Nash bargaining solution, preserving diversity under stated conditions.
-
Enabling Federated Inference via Unsupervised Consensus Embedding
CE-FI maps heterogeneous model representations to a shared embedding space via unsupervised training on unlabeled data, enabling privacy-preserving federated inference that outperforms solo models on image classification benchmarks.
-
A Dual Perspective on Synthetic Trajectory Generators: Utility Framework and Privacy Vulnerabilities
A new framework evaluates utility of synthetic mobility trajectories while a membership inference attack reveals privacy vulnerabilities in generative models thought to be safe.
-
Embedding Arithmetic: A Lightweight, Tuning-Free Framework for Post-hoc Bias Mitigation in Text-to-Image Models
Embedding Arithmetic performs vector operations in the embedding space of T2I models to mitigate bias at inference time, outperforming baselines on diversity while preserving coherence via a new Concept Coherence Score.
-
FatigueFusion: Latent Space Fusion for Fatigue-Driven Motion Synthesis
FatigueFusion fuses fatigue features in latent space using algorithmic, data-driven, and PINN modules to synthesize novel fatigued motions from non-fatigued joint sequences in an end-to-end pipeline.
-
gen2seg: Generative Models Enable Generalizable Instance Segmentation
Finetuning generative models on limited instance segmentation data produces zero-shot generalization to unseen object categories and styles, matching or exceeding supervised baselines like SAM on ambiguous boundaries.
-
"Noisier" Noise Contrastive Eestimation is (Almost) Maximum Likelihood
Scaling noise magnitude in NCE aligns gradients with MLE, enabling a practical approximation that improves performance on CIFAR-10 and ImageNet image modeling with fewer training steps.
-
MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework
MetaGPT embeds human SOPs into LLM prompts to create role-specialized agent teams that produce more coherent solutions on collaborative software engineering tasks than prior chat-based multi-agent systems.
-
VideoGPT: Video Generation using VQ-VAE and Transformers
VideoGPT generates competitive natural videos by learning discrete latents with VQ-VAE and modeling them autoregressively with a transformer.
-
Dual Adversarial Semantics-Consistent Network for Generalized Zero-Shot Learning
DASCN uses a unified primal-dual GAN architecture to generate semantics-consistent visual features for generalized zero-shot learning, claiming state-of-the-art gains.
-
Dual Adversarial Learning with Attention Mechanism for Fine-grained Medical Image Synthesis
Dual-discriminator GAN with adversarial attention improves fine-grained medical image synthesis, especially in hard-to-generate tumor regions, and outperforms prior methods on brain tumor and CT-to-MRI tasks.
-
RED: A ReRAM-based Deconvolution Accelerator
RED introduces pixel-wise mapping and zero-skipping dataflow for ReRAM deconvolution acceleration, reporting 1.15x-3.69x speedup and 8%-88.36% energy reduction versus prior ReRAM accelerators.
-
A Halo Merger Tree Generation and Evaluation Framework
A GAN framework is trained on EAGLE simulation merger trees to generate new realistic trees for semi-analytic galaxy models at modest computational cost.
-
Demystifying MMD GANs
MMD GANs have unbiased critic gradients but biased generator gradients from sample-based learning, and the Kernel Inception Distance provides a practical new measure for GAN convergence and dynamic learning rate adaptation.
-
APE: Agentic Prompt Enhancer for Image Generation and Editing
APE post-trains small language models as single-agent or multi-agent prompt enhancers that improve visual alignment on image generation and editing benchmarks without altering the downstream visual model.
-
Functional MRI Time Series Generation via Wavelet-Based Image Transform and Spectral Flow Matching for Brain Disorder Identification
DSFM generates synthetic fMRI time series via DWT-DCT dual transforms and spectral flow matching to improve downstream brain network classification.
-
Adversarial Error Correction for Visual Autoregressive Generation
AID-VAR attaches an adversarial discriminator and lightweight guidance injector to frozen VAR backbones to diagnose and correct fidelity gaps across scales, reporting 16% FID gains with 3% added parameters.
-
Are Candidate Models Really Needed for Active Learning?
Active learning with randomly initialized models achieves comparable results to traditional candidate-model methods, with low-confidence sampling proving most effective.
-
Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling
Visual generation models are evolving from passive renderers to interactive agentic world modelers, but current systems lack spatial reasoning, temporal consistency, and causal understanding, with evaluations overemphasizing perceptual quality.
-
ACPO: Anchor-Constrained Perceptual Optimization for Diffusion Models with No-Reference Quality Guidance
ACPO uses anchor-based regularization with NR-IQA guidance to enable stable perceptual quality improvements in diffusion model fine-tuning.
-
Improving Diversity in Black-box Few-shot Knowledge Distillation
An adaptive high-confidence image selection scheme during GAN training expands diversity in the distillation set for black-box few-shot KD and yields SOTA student accuracy on seven image datasets.
-
Physics-informed, Generative Adversarial Design of Funicular Shells
A GAN guided by a surrogate membrane-factor predictor generates 64-by-64 height-map shell shapes, but the physical-constraint implementation and independent validation are insufficient to establish compression-only funicular shells.
-
A Geometric Algebra-informed NeRF Framework for Generalizable Wireless Channel Prediction
GAI-NeRF combines geometric algebra attention and an adaptive ray tracing module inside a NeRF model to deliver more accurate and generalizable wireless channel predictions across varied indoor environments.
-
Unsupervised Detection of Spatiotemporal Anomalies in PMU Data Using Transformer-Based BiGAN
T-BiGAN integrates window-attention Transformers in a BiGAN to achieve ROC-AUC 0.95 and average precision 0.996 for unsupervised spatiotemporal anomaly detection in PMU data.
-
Quantum generative modeling for financial time series with temporal correlations
QGANs with quantum generators and classical discriminators generate financial time series matching target distributions and desired temporal correlations, with quality varying by circuit depth, bond dimension, and simulation method.
-
CCNETS: A Modular Causal Learning Framework for Pattern Recognition in Imbalanced Datasets
CCNETS is a new modular causal framework using three cooperative modules and a Zoint mechanism to align synthetic data generation with classifier needs on imbalanced pattern recognition tasks.
-
Synthetic Augmentation and Feature-based Filtering for Improved Cervical Histopathology Image Classification
cGAN data augmentation with feature-based filtering improves ResNet18 CIN grading accuracy from 66.3% to 71.7% on segmented epithelium patches.