Show HN: I invented a new generative model and got accepted to ICLR

(discrete-distribution-networks.github.io)

238 points | by diyer22 5 hours ago

14 comments

  • michaeldoron 55 minutes ago
    Very impressive to see a single author paper in ICLR, especially for an innovative method. Well done!
  • f_devd 3 hours ago
    Pretty interesting architecture, seems very easy to debug, but as a downside you effectively discard K-1 computations at each layer since it's using a sampler rather than a MoE-style router.

    The best way I can summarize it is a Mixture-of-Experts combined with an 'x0-target' latent diffusion model. The main innovation is the guided sampler (rather than router) & split-and-prune optimizer; making it easier to train.

    • yorwba 3 hours ago
      Since the sampling probability is 1/K independent of the input, you don't need to compute K different intermediate outputs at each layer during inference, you can instead decide ahead of time which of the outputs you want to use and only compute that one.

      (This is mentioned in Q1 in the "Common Questions About DDN" section at the bottom.)

      • kevmo314 13 minutes ago
        This is a very clever insight, nice work!
  • CuriouslyC 18 minutes ago
    Pretty interesting. I was just doing research on diffusion using symbolic transform matrices to try and parallelize a deep graph reactive system a few days ago, seems to be a general direction that people are going, I wouldn't be surprised to see diffusion adjacent models take over for codegen in the next year or two.
  • Lerc 55 minutes ago
    It's not often you read a title like that and expect it to pan out, but from a quick browse, it looks pretty good.

    Now I just need a time-turner.

  • moconnor 1 hour ago
    Super cool, I spent a lot of time playing with representation learning back in the day and the grids of MNIST digits took me right back :)

    A genuinely interesting and novel approach, I'm very curious how it will perform when scaled up and applied to non-image domains! Where's the best place to follow your work?

  • FitchApps 2 hours ago
    Can you train this model to detect objects (e.g detect a fish in the picture)?
    • diyer22 2 hours ago
      I believe DDN is exceptionally well-suited to the “generative models for discriminative tasks” paradigm for object detection.

      Much like DiffusionDet, which applies diffusion models to detection, DDN can adopt the same philosophy. I expect DDN to offer several advantages over diffusion-based approaches: - Single forward pass to obtain results, no iterative denoising required. - If multiple samples are needed (e.g., for uncertainty estimation), DDN can directly produce multiple outputs in one forward pass. - Easy to impose constraints during generation due to DDN's Zero-Shot Conditional Generation capability. - DDN supports more efficient end-to-end optimization, thus more suitable for integration with discriminative models and reinforcement learning.

  • gurtinator 1 hour ago
    How did this get accepted without any baseline comparisons? They should have compared this to VQ-VAE, diffusion inpainting and a lot more.
    • diyer22 58 minutes ago
      I believe it is the novelty. Here I would like to quote Reviewer r4YK’s original words:

      > Many high rated papers would have been done by someone else if their authors never published them or were rejected. However, if this paper is not published, it is not likely that anyone would come up with this approach. This is real publication value. I am reminding again the original diffusion paper from 2015 (Sohl-Dickstein) that was almost not noticed for 5 years. Had it not been published, would we have had the amazing generative models we have today?

      Cite from: https://openreview.net/forum?id=xNsIfzlefG&noteId=Dl4bXmujh1

      Besides, we compared DDN with other approaches in the Table 1 of original paper, including VQ-VAE.

  • GaggiX 1 hour ago
    It's so cool to see the hierarchical generation of the model, on their Github page they have one with L=4: https://discrete-distribution-networks.github.io/img/tree-la...

    The one shown on their page is L=3.

  • VoidWhisperer 4 hours ago
    I don't have a super deep understanding of the underlying algorithms involved, but going off the demo and that page, is this mainly a model for image related tasks, or could it also be trained to do things like what GPT/Claude/etc does (chat conversations)?
    • diyer22 3 hours ago
      Yes, it's absolutely possible—just like how diffusion LLMs work, we can do the same with DDN LLMs.

      I made an initial attempt to combine [DDN with GPT](https://github.com/Discrete-Distribution-Networks/Discrete-D...), aiming to remove tokenizers and let LLMs directly model binary strings. In each forward pass, the model adaptively adjusts the byte length of generated content based on generation difficulty (naturally supporting speculative sampling).

      • vintermann 23 minutes ago
        This is what I find most impressive, that it's a natural hierarchial method which seems so general, yet is actually quite competitive. I feel like the machine learning community has been looking for that for a long time. Non-generative uses (like hierarchial embeddings, maybe? Making Dewey's decimal like embeddings for anything!) are even more exciting.
        • diyer22 12 minutes ago
          Exactly! The paragraph on Efficient Data Compression Capability in the original paper also highlights:

          > To our knowledge, Taiji-DDN is the first generative model capable of directly transforming data into a semantically meaningful binary string which represents a leaf node on a balanced binary tree.

          This property excites me just as much.

    • booli 4 hours ago
  • p1esk 3 hours ago
    How does it compare to state of the art models? Does it scale?
    • diyer22 3 hours ago
      The first version of DDN was developed in less than three months, almost entirely by one person. Consequently, the experiments were preliminary and the results far from SoTA.

      The current goal in research is scaling up. Here are some thoughts in blog about future directions: https://github.com/Discrete-Distribution-Networks/Discrete-D...

  • Der_Einzige 2 hours ago
    Wtf, iclr reviews are happening right now. Did you get accepted into a workshop? How do you know it’s been accepted?
    • albertzeyer 2 hours ago
      ICLR 2026 reviews are happening now (or soon). This paper here was accepted at ICLR 2025.
  • nvr219 2 hours ago
    Congrats!! Very cool.
    • curtistyr 1 hour ago
      I've been thinking about this too—how different DDN is from other generative models. The idea of generating multiple outputs at once in a single pass sounds like it could really speed things up, especially for tasks where you need a bunch of samples quickly. I'm curious how this compares to something like GANs, which can also generate multiple samples but often struggle with mode collapse.

      The zero-shot conditional generation part is wild. Most methods rely on gradients or fine-tuning, so I wonder what makes DDN tick there. Maybe the tree structure of the latent space helps navigate to specific conditions without needing retraining? Also, I'm intrigued by the 1D discrete representation—how does that even work in practice? Does it make the model more interpretable?

      The Split-and-Prune optimizer sounds new—I'd love to see how it performs against Adam or SGD on similar tasks. And the fact that it's fully differentiable end-to-end is a big plus for training stability.

      I also wonder about scalability—can this handle high-res images without blowing up computationally? The hierarchical approach seems promising, but I'm not sure how it holds up when moving from simple distributions to something complex like natural images.

      Overall though, this feels like one of those papers that could really shift the direction of generative models. Excited to dig into the code and see what kind of results people get with it!

  • serf 3 hours ago
    isn't this kind of like an 80% vq-vae?
    • diyer22 2 hours ago
      No, DDN and VQ-VAE are clearly different.

      Similarities: - Both map data to a discrete latent space.

      Differences: - VQ-VAE needs an external prior over code indices (e.g. PixelCNN or a hierarchical prior) to model distribution. DDN builds its own hierarchical discrete distribution and can even act as the prior for a VQ-VAE-like system. - DDN’s K outputs are features that change with the input; VQ-VAE’s codebook is a set of independent parameters (embeddings) that remain fixed regardless of the input. - VQ-VAE produces a 2-D grid of code indices; DDN yields a 1-D/tree-structured latent. - VQ-VAE needs Straight-Through Estimator. - DDN supports zero-shot conditional generation.

      So I’d call them complementary rather than “80 % the same.” (See the paper’s “Connections to VQ-VAE.”)

  • tensorlibb 1 hour ago
    [dead]