In the rapidly evolving field of deep learning, the demand for models that are both expressive and computationally efficient has never been more critical. This paper introduces Orchid, a novel architecture designed to address the quadratic complexity of traditional attention mechanisms without compromising the ability to capture long-range dependencies and in-context learning. At the core of this architecture lies a new data-dependent global convolution layer, which contextually adapts its kernel conditioned on input sequence using a dedicated conditioning neural network. We design two simple conditioning networks that maintain shift equivariance in our data-dependent convolution operation. The dynamic nature of the proposed convolution kernel grants Orchid high expressivity while maintaining quasilinear scalability for long sequences. We evaluate the proposed model across multiple domains, including language modeling and image classification, to highlight its performance and generality. Our experiments demonstrate that this architecture not only outperforms traditional attention-based architectures such as BERT and Vision Transformers with smaller model sizes, but also extends the feasible sequence length beyond the limitations of the dense attention layers. This achievement represents a significant step towards more efficient and scalable deep learning models for sequence modeling.

Neural Information Processing Systems http://nips.cc/

Variational inference often struggles with the posterior geometry exhibited by complex hierarchical Bayesian models. Recent advances in flow-based variational families and Variationally Inferred Parameters (VIP) each address aspects of this challenge, but their formal relationship is unexplored. Here, we prove that the combination of VIP and a full-rank Gaussian can be represented exactly as a forward autoregressive flow augmented with a translation term and input from the model's prior. Guided by this theoretical insight, we introduce the Model-Informed Flow (MIF) architecture, which adds the necessary translation mechanism, prior information, and hierarchical ordering. Empirically, MIF delivers tighter posterior approximations and matches or exceeds state-of-the-art performance across a suite of hierarchical and non-hierarchical benchmarks.

In the rapidly evolving field of deep learning, the demand for models that are both expressive and computationally efficient has never been more critical. This paper introduces Orchid, a novel architecture designed to address the quadratic complexity of traditional attention mechanisms without compromising the ability to capture long-range dependencies and in-context learning. At the core of this architecture lies a new data-dependent global convolution layer, which contextually adapts its kernel conditioned on input sequence using a dedicated conditioning neural network. We design two simple conditioning networks that maintain shift equivariance in our data-dependent convolution operation. The dynamic nature of the proposed convolution kernel grants Orchid high expressivity while maintaining quasilinear scalability for long sequences. We evaluate the proposed model across multiple domains, including language modeling and image classification, to highlight its performance and generality. Our experiments demonstrate that this architecture not only outperforms traditional attention-based architectures such as BERT and Vision Transformers with smaller model sizes, but also extends the feasible sequence length beyond the limitations of the dense attention layers. This achievement represents a significant step towards more efficient and scalable deep learning models for sequence modeling.

Deep generative models are known to produce undesirable samples such as harmful content. Traditional mitigation methods include re-training from scratch, filtering, or editing; however, these are either computationally expensive or can be circumvented by third parties. In this paper, we take a different approach and study how to post-edit an already-trained conditional generative model so that it redacts certain conditionals that will, with high probability, lead to undesirable content. This is done by distilling the conditioning network in the models, giving a solution that is effective, efficient, controllable, and universal for a class of deep generative models. We conduct experiments on redacting prompts in text-to-image models and redacting voices in text-to-speech models. Our method is computationally light, leads to better redaction quality and robustness than baseline methods while still retaining high generation quality.

We introduce a new architecture called a conditional invertible neural network (cINN), and use it to address the task of diverse image-to-image translation for natural images. This is not easily possible with existing INN models due to some fundamental limitations. The cINN combines the purely generative INN model with an unconstrained feed-forward network, which efficiently preprocesses the conditioning image into maximally informative features. All parameters of a cINN are jointly optimized with a stable, maximum likelihood-based training procedure. Even though INN-based models have received far less attention in the literature than GANs, they have been shown to have some remarkable properties absent in GANs, e.g. apparent immunity to mode collapse. We find that our cINNs leverage these properties for image-to-image translation, demonstrated on day to night translation and image colorization. Furthermore, we take advantage of our bidirectional cINN architecture to explore and manipulate emergent properties of the latent space, such as changing the image style in an intuitive way.

Semantic inpainting is the task of inferring missing pixels in an image given surrounding pixels and high level image semantics. Most semantic inpainting algorithms are deterministic: given an image with missing regions, a single inpainted image is generated. However, there are often several plausible inpaintings for a given missing region. In this paper, we propose a method to perform probabilistic semantic inpainting by building a model, based on PixelCNNs, that learns a distribution of images conditioned on a subset of visible pixels. Experiments on the MNIST and CelebA datasets show that our method produces diverse and realistic inpaintings. Further, our model also estimates the likelihood of each sample which we show correlates well with the realism of the generated inpaintings.

Google Brain's latest software can create sharpen images from a pixelated source. The system combines two neural networks and machine learning to guess what details lay hidden in the blurry picture. Once the system is fed an 8 x 8 pixelated image, the networks search for high-resolution images that it believes matches the source's content - and adds the missing details. The system combines two neural networks and machine learning to guess what details lay hidden in the blurry picture. Once the system is fed an 8 x 8 pixelated image, the networks search for high-resolution images that it believes matches the source's content The team at Google Brain has developed a system that is capable of making out details of a pixelated source.