REVIEW 40 cited by
Masked Diffusion Models are Secretly Time-Agnostic Masked Models and Exploit Inaccurate Categorical Sampling
Not yet reviewed by Pith; the record is open.
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
Masked Diffusion Models are Secretly Time-Agnostic Masked Models and Exploit Inaccurate Categorical Sampling
read the original abstract
Masked diffusion models (MDMs) have emerged as a popular research topic for generative modeling of discrete data, thanks to their superior performance over other discrete diffusion models, and are rivaling the auto-regressive models (ARMs) for language modeling tasks. The recent effort in simplifying the masked diffusion framework further leads to alignment with continuous-space diffusion models and more principled training and sampling recipes. In this paper, however, we reveal that both training and sampling of MDMs are theoretically free from the time variable, arguably the key signature of diffusion models, and are instead equivalent to masked models. The connection on the sampling aspect is drawn by our proposed first-hitting sampler (FHS). Specifically, we show that the FHS is theoretically equivalent to MDMs' original generation process while significantly alleviating the time-consuming categorical sampling and achieving a 20$\times$ speedup. In addition, our investigation raises doubts about whether MDMs can truly beat ARMs in text generation. We identify, for the first time, an underlying numerical issue, even with the commonly used 32-bit floating-point precision, which results in inaccurate categorical sampling. We show that it lowers the effective temperature both theoretically and empirically, and the resulting decrease in token diversity makes previous evaluations, which assess the generation quality solely through the incomplete generative perplexity metric, somewhat unfair.
Forward citations
Cited by 40 Pith papers
-
Large Language Diffusion Models
LLaDA is a scalable diffusion-based language model that matches autoregressive LLMs like LLaMA3 8B on tasks and surpasses GPT-4o on reversal poem completion.
-
Set Diffusion: Interpolating Token Orderings Between Autoregression and Diffusion for Fast and Flexible Decoding
Set diffusion factorizes likelihood over arbitrary token sets and uses a set-causal diffusion architecture to support KV caching and any-order decoding, yielding improved speed-quality tradeoffs versus prior diffusion LMs.
-
Masked Language Flow Models
MLFMs combine masking with continuous flows to scale flow-based language models to reasoning and instruction-following tasks on GSM8K and MT-Bench.
-
Continuous Language Diffusion as a Decoder-Interface Problem
Continuous language diffusion works by entering high-margin decoder basins where frozen T5 embeddings recover 93-96% of native decisions and linear readouts reach 97.9% agreement, implying models should be evaluated a...
-
Learned Relay Representations for Forward-Thinking Discrete Diffusion Models
Learned Relay Representations enable masked diffusion models to propagate useful latent information across denoising steps, scaling to Fast-dLLM v2 to outperform supervised finetuning on coding tasks while cutting inf...
-
Uniform Diffusion Models Revisited: Leave-One-Out Denoiser and Absorbing State Reformulation
Uniform diffusion models rely on a leave-one-out denoiser rather than the usual denoising posterior, with exact conversions derived; an absorbing-state reformulation is introduced that matches or exceeds masked diffus...
-
Self-Distilled Trajectory-Aware Boltzmann Modeling: Bridging the Training-Inference Discrepancy in Diffusion Language Models
TABOM is a trajectory-aligned Boltzmann modeling framework that turns self-distilled inference paths into a pairwise ranking loss to close the training-inference gap in diffusion language models and expand their effec...
-
Simple Self-Conditioning Adaptation for Masked Diffusion Models
SCMDM is a post-training self-conditioning adaptation for masked diffusion models that reduces generative perplexity by nearly 50% on OWT and improves performance on images, molecules, and genomics.
-
Discrete Tilt Matching
Discrete Tilt Matching recasts dLLM fine-tuning as state-level matching of tilted local unmasking posteriors, producing a stable weighted cross-entropy loss that improves Sudoku and Countdown performance when applied ...
-
Discrete Tilt Matching
DTM recasts dLLM fine-tuning as weighted cross-entropy matching of tilted local posteriors, with demonstrated gains on Sudoku and math tasks.
-
Coevolving Representations in Joint Image-Feature Diffusion
CoReDi coevolves semantic representations with the diffusion model via a jointly learned linear projection stabilized by stop-gradient, normalization, and regularization, yielding faster convergence and higher sample ...
-
DMax: Aggressive Parallel Decoding for dLLMs
DMax uses On-Policy Uniform Training and Soft Parallel Decoding to enable aggressive parallelism in dLLMs, raising TPF on GSM8K from 2.04 to 5.47 and on MBPP from 2.71 to 5.86 while preserving accuracy.
-
MemDLM: Memory-Enhanced DLM Training
MemDLM embeds a simulated denoising trajectory into DLM training via bi-level optimization, creating a parametric memory that improves convergence and long-context performance even when the memory is dropped at test time.
-
Flow Map Language Models: One-step Language Modeling via Continuous Denoising
Continuous flow language models match discrete diffusion baselines and their distilled one-step flow map versions exceed 8-step discrete diffusion quality on LM1B and OWT.
-
Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding
Fast-dLLM adds reusable KV cache blocks and selective parallel decoding to diffusion LLMs, closing most of the speed gap with autoregressive models without retraining.
-
From Interface to Inference: Eliciting Any-Order Inference from Any-Order Models
Masked diffusion language models fail to exploit their any-order interface because of positional uncertainty; insertion-based (FlexMDM) and latent-segment (LatentMDM) variants recover distinct any-order inference beha...
-
Concurrent Image Understanding and Generation: Self-Correcting Coupled Markov Jump Processes
CO2Jump couples text and image denoising through cross-modal attention and remasking, achieving best joint accuracy on three concurrent-generation tasks.
-
Accelerating Discrete Diffusion Models with Parallel-In-Time Sampling
A parallel-in-time τ-leaping sampler for absorbing discrete diffusion models is introduced, with an exponential-factorial convergence proof and empirical speedups of 7-9× on synthetic tasks and 1.45-1.86× on image/tex...
-
Posterior Refinement: Fast Language Generation via Any-Order Flow Maps
FMLM+ with Posterior Refinement bridges masked diffusion and flow map models to match discrete baseline quality in language generation using 32x fewer neural function evaluations via posterior scoring and refinement.
-
CaRE Compute-aware Remasking Evaluation Protocol for Masked Diffusion Language Models
CaRE shows that uncontrolled temperature and compute explain most MAUVE differences between MDLM remasking strategies, with high-entropy remasking reducing MAUVE under stochastic unmasking.
-
NAVIRA: Decoupled Stochastic Remasking for Masked Diffusion Language Models
NAVIRA decouples quality scoring from regeneration via stochastic remasking in masked diffusion LMs, improving fluency and LLM-judge scores on a 170M model.
-
Fixed-Point Masked Generative Modeling
FP-MGMs with consistency loss and three-state reuse (CoFRe) reduce parameters by up to 38.8% and improve low-budget perplexity and FID versus standard masked generative models on text and images.
-
dMoE: dLLMs with Learnable Block Experts
dMoE aggregates token expert distributions to block level in dLLMs, cutting unique experts from 69.5 to 14.6, memory by 76-80%, and latency by 1.14-1.66x while retaining 99.11% performance.
-
Learned Relay Representations for Forward-Thinking Discrete Diffusion Models
Learned Relay Representations add a differentiable per-token channel to masked diffusion models so they can propagate latent information across iterative denoising steps, yielding better coding performance and up to 3...
-
Self-Distilled Trajectory-Aware Boltzmann Modeling: Bridging the Training-Inference Discrepancy in Diffusion Language Models
TABOM models inference unmasking preferences as a Boltzmann distribution over predictive entropies and derives a ranking loss to align DLM training with observed trajectories, yielding gains in new domains and reduced...
-
Continuous Latent Diffusion Language Model
Cola DLM proposes a hierarchical latent diffusion model that learns a text-to-latent mapping, fits a global semantic prior in continuous space with a block-causal DiT, and performs conditional decoding, establishing l...
-
Token Time Continuous Diffusion for Language Modeling
A continuous diffusion language model where each token denoises at its own rate—sure tokens first—improves few-step generation over discrete samplers and roughly matches global-time continuous models.
-
Simple Self-Conditioning Adaptation for Masked Diffusion Models
SCMDM adapts trained masked diffusion models to condition denoising steps on their own prior clean predictions, cutting generative perplexity nearly in half on open-web text while improving discretized image, molecule...
-
Generative Frontiers: Why Evaluation Matters for Diffusion Language Models
Generative perplexity and entropy are shown to be the two additive components of KL divergence to a reference distribution, motivating generative frontiers as a principled evaluation method for diffusion language models.
-
Is Your Diffusion Sampler Actually Correct? A Sampler-Centric Evaluation of Discrete Diffusion Language Models
Under an exact oracle denoiser derived from a ground-truth Markov chain, SEDD, MDLM, LLaDA, and ReMDM samplers show large transition-level error at few diffusion steps; GenPPL/MAUVE can improve or stay flat while tran...
-
Flow Map Language Models: One-step Language Modeling via Continuous Denoising
Continuous flows on token embeddings with flow-map distillation produce one-step language models whose quality exceeds recent 8-step discrete diffusion baselines on LM1B and OpenWebText.
-
CANDI: Hybrid Discrete-Continuous Diffusion Models
CANDI combines masked and Gaussian corruption in one noising process, letting discrete diffusion models use continuous gradients for joint updates and guidance.
-
Fine-Tuning Masked Diffusion for Provable Self-Correction
PRISM fine-tunes any masked diffusion model with a binary-cross-entropy loss so its new head provably estimates per-token quality p(x_i=y_i|y⊕m_i) and can remask low-quality tokens at inference.
-
Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models
TraceRL aligns the RL objective with the DLM's actual step-by-step decoding, producing TraDo-4B/8B models that beat autoregressive baselines on math reasoning.
-
Any-Order Flexible Length Masked Diffusion
FlexMDM is a discrete diffusion model that provably supports any-order generation over variable-length sequences by learning an insertion expectation alongside the unmasking posterior, validated by length-fidelity, ma...
-
Diffusion Language Models Know the Answer Before Decoding
DLMs show early answer convergence allowing Prophet to cut decoding steps by up to 3.4x on LLaDA-8B and Dream-7B while keeping output quality.
-
LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning
LLaDA-V is a diffusion-based multimodal large language model that reaches competitive or state-of-the-art results on visual instruction tasks while using a non-autoregressive architecture.
-
TACG: Trajectory-Aware Commit Gating for Diffusion Language Model Decoding
Trajectory-aware commit gating (TILG + History Gate + capped extra promotion) improves or preserves DLLM accuracy while reducing steps and raising tokens-per-forward without retraining.
-
Explorative Modeling: Unlocking a Third Pretraining Axis and End-to-End Generation
Training on the best of K generated candidates improves image, video, and language generative models, with the reported gains growing with scale and enabling single-pass end-to-end generation.
-
DMax: Aggressive Parallel Decoding for dLLMs
DMax enables faster parallel decoding in diffusion language models by using on-policy training to recover from errors and soft embedding interpolations for iterative revision, boosting tokens per forward pass roughly ...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.