Tagged articles
generative modeling
16 articles · Page 1 of 1
Tadpole introduces an autoencoder-based foundation model for 3D partial differential equations, pretrained via reconstruction on massive online-generated simulation data (202 TB equivalent) without storing it, enabling zero-shot reconstruction, dynamics prediction via lightweight adaptation, and generation across diverse 3D physical systems and resolutions.
3D PDEZero-shot Generalizationautoencoder
0 likes · 14 min read

SCoPE (Sightline‑Coordinate Positional Encoding) replaces the traditional (u,v,t) pixel‑grid coordinates in video DiT models with camera‑ray coordinates injected into attention, achieving near‑zero parameter overhead while markedly improving camera control, cross‑view consistency, and revisit consistency across model scales and stylized inputs.
AttentionComputer VisionRay‑Space Positional Encoding
0 likes · 9 min read

By swapping the traditional MSE loss for a perceptual loss in Flow Matching training, the authors enable high‑quality diffusion generation in only 4–8 inference steps—down from 35–50—without teacher models, distribution or trajectory distillation, and they substantiate the claim with extensive experiments and a new distribution‑distance metric.
Computer VisionDistribution DistanceFlow Matching
0 likes · 7 min read

Diffusion models have set the state‑of‑the‑art in image, video, and audio generation, yet their training objective admits a unique closed‑form solution that merely memorizes training data; this tutorial examines why they still generalize by exploring score smoothing, architectural inductive bias, training dynamics, and data geometry, all illustrated with hands‑on Jupyter notebooks.
CVPR 2026Generalizationdata geometry
0 likes · 2 min read

The article traces generative learning from its probabilistic roots in Bayesian classification, through Gaussian mixture models, hidden Markov models, N‑gram and neural language models, to attention mechanisms, Transformers and GPT, highlighting how each innovation expanded the ability to model data‑generating processes.
BayesianGPTGaussian Mixture
0 likes · 26 min read

Xiaomi Auto introduces a novel world‑model framework that tightly couples 3D reconstruction and generative prediction, delivering state‑of‑the‑art performance on Waymo and nuScenes benchmarks while enabling high‑fidelity, long‑duration video synthesis for autonomous‑driving scenarios.
3D reconstructionAutonomous DrivingBenchmark SOTA
0 likes · 10 min read

This survey introduces vision world models as a central driver for AI to learn physical and causal dynamics directly from visual data, presents a unified "representation‑learning‑simulation" framework, categorises four major technical routes, outlines evaluation metrics and datasets, and proposes a 3R roadmap for the next generation of world models.
Evaluation MetricsFuture DirectionsVision World Models
0 likes · 15 min read

Diffolio introduces a diffusion‑model‑based approach that directly learns a pseudo‑optimal portfolio distribution conditioned on user risk preferences, generating diverse high‑quality portfolios and outperforming classic and recent baselines on six real‑world market datasets, with annualized returns improving up to 12.1 percentage points.
Quantitative Financediffusion modelfinancial AI
0 likes · 22 min read

The article analyzes the evolution from traditional modular recommendation models to a generative Large User Model (LUM), detailing its three‑stage paradigm, tokenization, training objectives, scaling‑law findings, offline and online experiments, and the AI‑infra innovations that enabled a 4.5% CTR lift in production.
CTR predictionScaling Lawsgenerative modeling
0 likes · 18 min read

A 17‑year‑old high‑school student from Anhui Ansheng School led the first‑author CVPR 2026 paper "CraftMesh," a novel 3D mesh editing framework that combines image editing, mesh generation, and Poisson seamless fusion, achieving superior quantitative metrics and showcasing the growing impact of young researchers in top AI conferences.
3D mesh generationCVPRComputer Vision
0 likes · 7 min read


AIWalkerMar 4, 2026 · Artificial Intelligence
The paper introduces Drifting Models, a new generative paradigm that moves the distribution evolution to the training phase, achieving true one‑step (1‑NFE) generation with state‑of‑the‑art ImageNet FID scores of 1.54 in latent space and 1.61 in pixel space, while eliminating the need for distillation or classifier‑free guidance.
Drifting ModelsImageNetdiffusion
0 likes · 24 min read

MIT’s SpectroGen model uses physics‑informed generative AI to convert a single spectral modality into high‑fidelity cross‑modal spectra, achieving up to 99% correlation with experimental data, dramatically reducing the cost and time of material spectroscopy while preserving detailed spectral features.
AIVariational AutoEncodergenerative modeling
0 likes · 9 min read

The LaVin-DiT paper presents a large vision diffusion transformer that integrates a spatio‑temporal variational auto‑encoder, a joint diffusion transformer with full‑sequence joint attention, and 3D rotary position encoding to enable unified, efficient multi‑task generation for images and videos, and details its training via flow‑matching and experimental results.
3D RoPEComputer VisionJoint Diffusion Transformer
0 likes · 12 min read

The OneRec system from Kuaishou replaces traditional cascade recommendation pipelines with an encoder‑decoder architecture, leverages reward‑based preference alignment via reinforcement learning, achieves ten‑fold FLOPs gains, cuts operational costs by 90%, and delivers significant user‑engagement improvements across short‑video and local‑service scenarios.
KuaishouOneRecRecommendation System
0 likes · 17 min read


AntTechJun 4, 2025 · Artificial Intelligence
This article presents the LLaDA series of diffusion‑based large language models, explains how their generative‑modeling principle yields language intelligence comparable to autoregressive models, and details the multimodal LLaDA‑V architecture, training methods, experimental results, and broader implications for AI research.
diffusion modelsgenerative modelinginstruction following
0 likes · 10 min read

Flow Matching reformulates generative modeling by learning a time‑dependent vector field that deterministically transports Gaussian noise to data, using a neural network trained with an analytically derived L2 loss, yielding simpler training, faster convergence, and deterministic sampling that matches or exceeds diffusion model quality.
AIFlow Matchingcontinuous normalizing flow
0 likes · 13 min read
