# MCLR: Improving Conditional Modeling via Inter-Class Likelihood-Ratio Maximization and Unifying Classifier-Free Guidance with Alignment Objectives

arXiv: https://arxiv.org/abs/2603.22364 · HTML: https://mlsanity.com/paper/2603.22364/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models.
- Builds on 17 of its 52 references that are in the map.
- First posted on arXiv 23 Mar 2026.

## Placed in

- Diffusion models / [Diffusion foundations and sampling](https://mlsanity.com/diffusion/diffusion-foundations-and-sampling/)

## Builds on (most cited first, 17 of 17)

- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Hierarchical Text-Conditional Image Generation with CLIP Latents](https://mlsanity.com/paper/2204.06125/) (2022)
- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [Classifier-Free Guidance is a Predictor-Corrector](https://mlsanity.com/paper/2408.09000/) (2024)
- [Diffusion Guidance Is a Controllable Policy Improvement Operator](https://mlsanity.com/paper/2505.23458/) (2025)
- [Diffusion Models without Classifier-free Guidance](https://mlsanity.com/paper/2502.12154/) (2025)
- [Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator](https://mlsanity.com/paper/2503.01103/) (2025)
- [Aligning Text to Image in Diffusion Models is Easier Than You Think](https://mlsanity.com/paper/2503.08250/) (2025)
- [Stage-wise Dynamics of Classifier-Free Guidance in Diffusion Models](https://mlsanity.com/paper/2509.22007/) (2025)
- [Classifier-Free Guidance: From High-Dimensional Analysis to Generalized Guidance Forms](https://mlsanity.com/paper/2502.07849/) (2025)
- [Inference-Time Alignment Control for Diffusion Models with Reinforcement Learning Guidance](https://mlsanity.com/paper/2508.21016/) (2025)
- [No Alignment Needed for Generation: Learning Linearly Separable Representations in Diffusion Models](https://mlsanity.com/paper/2509.21565/) (2025)
- [Diffusion Blend: Inference-Time Multi-Preference Alignment for Diffusion Models](https://mlsanity.com/paper/2505.18547/) (2025)
- [Training Class-Imbalanced Diffusion Model Via Overlap Optimization](https://mlsanity.com/paper/2402.10821/) (2024)
- [Emergence of Distortions in High-Dimensional Guided Diffusion Models](https://mlsanity.com/paper/2602.00716/) (2026)
- [Feature-guided score diffusion for sampling conditional densities](https://mlsanity.com/paper/2410.11646/) (2024)
- [Rethinking Preference Alignment for Diffusion Models with Classifier-Free Guidance](https://mlsanity.com/paper/2602.18799/) (2026)

## Built on by (most cited first, 2 of 2)

- [Evaluating the Representation Space of Diffusion Models via Self-Supervised Principles](https://mlsanity.com/paper/2606.09718/) (2026)
- [Scaling and Distilling Text Embeddings for Better Diffusibility](https://mlsanity.com/paper/2610.01016/) (2026)

## Abstract (arXiv)

Diffusion models achieve strong performance in generative modeling, but their success often relies heavily on classifier-free guidance (CFG), an inference-time heuristic that modifies the sampling trajectory. In theory, diffusion models trained with standard denoising score matching (DSM) should recover the target data distribution, raising two fundamental questions: (i) why is inference-time guidance necessary in practice, and (ii) can its underlying effect be internalized into a principled training objective? In this work, we argue that a key limitation of standard DSM is insufficient inter-class separation. To address this issue, we propose MCLR, an alignment objective that explicitly maximizes inter-class likelihood-ratios during training. Fine-tuning diffusion models with MCLR induces CFG-like improvements under standard sampling, substantially improving guidance-free conditional generation and narrowing the gap to inference-time CFG. Beyond these empirical benefits, we show theoretically that the CFG-guided score is exactly the optimal solution to a sample-adaptive weighted MCLR objective. This result connects CFG to alignment-based objectives, providing a mechanistic interpretation of CFG as an implicit inference-time contrastive alignment procedure.