Pith. sign in

REVIEW 15 cited by

Beyond Next-Token: Next-X Prediction for Autoregressive Visual Generation

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2502.20388 v2 pith:GGAQ7QBW submitted 2025-02-27 cs.CV

Sucheng Ren , Qihang Yu , Ju He , Xiaohui Shen , Alan Yuille , Liang-Chieh Chen This is my paper
classification cs.CV
keywords tokenpredictionbiasexposuretimesautoregressivediscreteentity
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

Autoregressive (AR) modeling, known for its next-token prediction paradigm, underpins state-of-the-art language and visual generative models. Traditionally, a ``token'' is treated as the smallest prediction unit, often a discrete symbol in language or a quantized patch in vision. However, the optimal token definition for 2D image structures remains an open question. Moreover, AR models suffer from exposure bias, where teacher forcing during training leads to error accumulation at inference. In this paper, we propose xAR, a generalized AR framework that extends the notion of a token to an entity X, which can represent an individual patch token, a cell (a $k\times k$ grouping of neighboring patches), a subsample (a non-local grouping of distant patches), a scale (coarse-to-fine resolution), or even a whole image. Additionally, we reformulate discrete token classification as continuous entity regression, leveraging flow-matching methods at each AR step. This approach conditions training on noisy entities instead of ground truth tokens, leading to Noisy Context Learning, which effectively alleviates exposure bias. As a result, xAR offers two key advantages: (1) it enables flexible prediction units that capture different contextual granularity and spatial structures, and (2) it mitigates exposure bias by avoiding reliance on teacher forcing. On ImageNet-256 generation benchmark, our base model, xAR-B (172M), outperforms DiT-XL/SiT-XL (675M) while achieving 20$\times$ faster inference. Meanwhile, xAR-H sets a new state-of-the-art with an FID of 1.24, running 2.2$\times$ faster than the previous best-performing model without relying on vision foundation modules (e.g., DINOv2) or advanced guidance interval sampling.

Discussion (0). Sign in to comment.

Forward citations

Cited by 15 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Revisiting Autoregressive Models for Generative Image Classification

    cs.CV 2026-03 accept novelty 6.5 of 10

    Order-marginalized any-order AR models (RandAR) outperform diffusion generative classifiers on ImageNet and OOD sets and match strong SSL models at far lower cost.

  2. MIMFlow: Integrating Masked Image Modeling with Normalizing Flows for End-to-End Image Generation

    cs.CV 2026-06 accept novelty 6.0 of 10

    End-to-end masked-image VAE plus normalizing flow yields FID 2.50 on ImageNet 256 with 128 tokens and higher linear-probe accuracy than unmasked counterparts.

  3. MIMFlow: Integrating Masked Image Modeling with Normalizing Flows for End-to-End Image Generation

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    MIMFlow uses a VAE on masked images to feed semantic latents to a normalizing flow while a decoder handles high-frequency details, reporting FID 2.50 and 71.3% linear probing on ImageNet 256x256 with 128 tokens.

  4. MIMFlow: Integrating Masked Image Modeling with Normalizing Flows for End-to-End Image Generation

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    MIMFlow is an end-to-end model that routes semantic latents through a normalizing flow while a decoder handles high-frequency pixels, reporting FID 2.50 and 71.3% linear probing accuracy on ImageNet 256x256 with 128 tokens.

  5. Data Provenance for Image Auto-Regressive Generation

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    A post-hoc detection framework exploits generation-induced patterns in autoregressive image outputs to enable provenance tracing across multiple IAR models without altering the generation process.

  6. IDEAL: In-DEpth ALignment Makes A Discrete Representation AutoEncoder

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    IDEAL improves discrete representation autoencoders by jointly aligning quantized tokens with shallow and deep VFM features, reporting 0.61 rFID on ImageNet and 1.89 gFID for autoregressive image generation.

  7. VPG: Visual Prefix Guidance for Autoregressive Image and Video Generation

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    VPG is a training-free inference-time guidance technique that improves autoregressive image and video generation by contrasting model outputs under generated versus corrupted prefixes to strengthen next-step support f...

  8. SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generation

    cs.CV 2026-05 conditional novelty 6.0 of 10

    A learned 32-channel semantic compression lets a normalizing flow reach gFID 1.65 on ImageNet 256×256, the best FID reported for flow-based image generation.

  9. SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generation

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    SRC-Flow compresses RAE features into a low-dimensional semantic space with a Semantic Representation Compressor, enabling normalizing flows to achieve SOTA gFID scores of 1.65 and 2.07 on ImageNet 256x256 and 512x512...

  10. SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generation

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    SRC-Flow compresses RAE features via a Semantic Representation Compressor into a low-dimensional space, enabling normalizing flows to reach gFID 1.65 on ImageNet 256x256 and 2.07 on 512x512 while retaining exact likelihoods.

  11. Protein Autoregressive Modeling via Multiscale Structure Generation

    cs.LG 2026-02 unverdicted novelty 6.0 of 10

    PAR is a multi-scale autoregressive transformer framework for protein backbone generation that uses coarse-to-fine prediction, noisy context learning, and flow-based decoding to achieve high-quality unconditional and ...

  12. Progressive Checkerboards for Autoregressive Multiscale Image Generation

    cs.CV 2026-02 conditional novelty 6.0 of 10

    A balanced multiscale checkerboard sampling order for autoregressive image generation allows large scale-up factors without quality loss, because only the total number of serial steps matters.

  13. Mirai: Autoregressive Visual Generation Needs Foresight

    cs.CV 2026-01 conditional novelty 6.0 of 10

    Mirai injects future-token foresight into autoregressive visual generators, accelerating convergence up to 10x and cutting ImageNet FID from 5.34 to 4.34.

  14. MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture

    cs.CV 2025-12 conditional novelty 6.0 of 10

    Training diffusion models on a mixture of higher-noise interpolations (MixFlow) improves generation FID across SiT, REPA, RAE and SD3.5, reaching ImageNet 256 gFID 1.43 after post-training.

  15. MEPA: Multi-Scale Representation Alignment for Visual Autoregressive Modeling with Mixture of Experts

    cs.CV 2026-07 unverdicted novelty 5.0 of 10

    MEPA adds token-routed MoE and residual self-supervised feature alignment to VAR models, reporting better FID on ImageNet 256x256 with half the training epochs and fewer parameters than dense baselines.

Pith tools