Vikram Voleti

Matt Deitke, Ruoshi Liu, Matthew Wallingford et al. · uw

Natural language processing and 2D vision models have attained remarkable proficiency on many tasks primarily by escalating the scale of training data. However, 3D vision tasks have not seen the same progress, in part due to the challenges of acquiring high-quality 3D data. In this work, we present Objaverse-XL, a dataset of over 10 million 3D objects. Our dataset comprises deduplicated 3D objects from a diverse set of sources, including manually designed objects, photogrammetry scans of landmarks and everyday items, and professional scans of historic and antique artifacts. Representing the largest scale and diversity in the realm of 3D datasets, Objaverse-XL enables significant new possibilities for 3D vision. Our experiments demonstrate the improvements enabled with the scale provided by Objaverse-XL. We show that by training Zero123 on novel view synthesis, utilizing over 100 million multi-view rendered images, we achieve strong zero-shot generalization abilities. We hope that releasing Objaverse-XL will enable further innovations in the field of 3D vision at scale.

Andreas Blattmann, Tim Dockhorn, Sumith Kulal et al.

We present Stable Video Diffusion - a latent video diffusion model for high-resolution, state-of-the-art text-to-video and image-to-video generation. Recently, latent diffusion models trained for 2D image synthesis have been turned into generative video models by inserting temporal layers and finetuning them on small, high-quality video datasets. However, training methods in the literature vary widely, and the field has yet to agree on a unified strategy for curating video data. In this paper, we identify and evaluate three different stages for successful training of video LDMs: text-to-image pretraining, video pretraining, and high-quality video finetuning. Furthermore, we demonstrate the necessity of a well-curated pretraining dataset for generating high-quality videos and present a systematic curation process to train a strong base model, including captioning and filtering strategies. We then explore the impact of finetuning our base model on high-quality data and train a text-to-video model that is competitive with closed-source video generation. We also show that our base model provides a powerful motion representation for downstream tasks such as image-to-video generation and adaptability to camera motion-specific LoRA modules. Finally, we demonstrate that our model provides a strong multi-view 3D-prior and can serve as a base to finetune a multi-view diffusion model that jointly generates multiple views of objects in a feedforward fashion, outperforming image-based methods at a fraction of their compute budget. We release code and model weights at https://github.com/Stability-AI/generative-models .

Vikram Voleti, Alexia Jolicoeur-Martineau, Christopher Pal

Video prediction is a challenging task. The quality of video frames from current state-of-the-art (SOTA) generative models tends to be poor and generalization beyond the training data is difficult. Furthermore, existing prediction frameworks are typically not capable of simultaneously handling other video-related tasks such as unconditional generation or interpolation. In this work, we devise a general-purpose framework called Masked Conditional Video Diffusion (MCVD) for all of these video synthesis tasks using a probabilistic conditional score-based denoising diffusion model, conditioned on past and/or future frames. We train the model in a manner where we randomly and independently mask all the past frames or all the future frames. This novel but straightforward setup allows us to train a single model that is capable of executing a broad range of video tasks, specifically: future/past prediction -- when only future/past frames are masked; unconditional generation -- when both past and future frames are masked; and interpolation -- when neither past nor future frames are masked. Our experiments show that this approach can generate high-quality frames for diverse types of videos. Our MCVD models are built from simple non-recurrent 2D-convolutional architectures, conditioning on blocks of frames and generating blocks of frames. We generate videos of arbitrary lengths autoregressively in a block-wise manner. Our approach yields SOTA results across standard video prediction and interpolation benchmarks, with computation times for training models measured in 1-12 days using $\le$ 4 GPUs. Project page: https://mask-cond-video-diffusion.github.io ; Code : https://github.com/voletiv/mcvd-pytorch

Jae Hyun Lim, Nikola B. Kovachki, Ricardo Baptista et al.

Diffusion models have recently emerged as a powerful framework for generative modeling. They consist of a forward process that perturbs input data with Gaussian white noise and a reverse process that learns a score function to generate samples by denoising. Despite their tremendous success, they are mostly formulated on finite-dimensional spaces, e.g., Euclidean, limiting their applications to many domains where the data has a functional form, such as in scientific computing and 3D geometric data analysis. This work introduces a mathematically rigorous framework called Denoising Diffusion Operators (DDOs) for training diffusion models in function space. In DDOs, the forward process perturbs input functions gradually using a Gaussian process. The generative process is formulated by a function-valued annealed Langevin dynamic. Our approach requires an appropriate notion of the score for the perturbed data distribution, which we obtain by generalizing denoising score matching to function spaces that can be infinite-dimensional. We show that the corresponding discretized algorithm generates accurate samples at a fixed cost independent of the data resolution. We theoretically and numerically verify the applicability of our approach on a set of function-valued problems, including generating solutions to the Navier-Stokes equation viewed as the push-forward distribution of forcings from a Gaussian Random Field (GRF), as well as volcano InSAR and MNIST-SDF.

Yiming Xie, Chun-Han Yao, Vikram Voleti et al.

We present Stable Video 4D (SV4D), a latent video diffusion model for multi-frame and multi-view consistent dynamic 3D content generation. Unlike previous methods that rely on separately trained generative models for video generation and novel view synthesis, we design a unified diffusion model to generate novel view videos of dynamic 3D objects. Specifically, given a monocular reference video, SV4D generates novel views for each video frame that are temporally consistent. We then use the generated novel view videos to optimize an implicit 4D representation (dynamic NeRF) efficiently, without the need for cumbersome SDS-based optimization used in most prior works. To train our unified novel view video generation model, we curate a dynamic 3D object dataset from the existing Objaverse dataset. Extensive experimental results on multiple datasets and user studies demonstrate SV4D's state-of-the-art performance on novel-view video synthesis as well as 4D generation compared to prior works.

Vikram Voleti, Christopher Pal, Adam Oberman

Generative models based on denoising diffusion techniques have led to an unprecedented increase in the quality and diversity of imagery that is now possible to create with neural generative models. However, most contemporary state-of-the-art methods are derived from a standard isotropic Gaussian formulation. In this work we examine the situation where non-isotropic Gaussian distributions are used. We present the key mathematical derivations for creating denoising diffusion models using an underlying non-isotropic Gaussian noise model. We also provide initial experiments with the CIFAR-10 dataset to help verify empirically that this more general modeling approach can also yield high-quality samples.

Ciara Rowles, Reshinth Adithyan, Nikhil Pinnaparaju et al.

We present Stable-Layers, a reinforcement learning framework that eliminates the need for paired supervision by fine-tuning a pretrained layer decomposition model using only feedback from a vision-language model (VLM). Starting from Qwen-Image-Layered, we apply Flow-GRPO with LoRA adaptation, sampling multiple candidate decompositions per image, scoring them with a VLM, and optimising the policy from group-relative advantages. The key challenge lies in designing a reliable reward signal: VLMs scoring samples in isolation tend to compress their judgements into a narrow band, leaving GRPO with little within-group variance to learn from. We address this with a two-stage evaluation pipeline that pairs structured per-sample scoring across five edit-centric criteria with a grid-based calibration step in which the VLM re-scores all candidates side-by-side. Stable-Layers produces decompositions with stronger layer separation, fewer blank or artifact-heavy layers, and lower per-layer reconstruction error on the Crello dataset compared to the base model.

Vikram Voleti, Boris N. Oreshkin, Florent Bocquelet et al.

Inverse Kinematics (IK) systems are often rigid with respect to their input character, thus requiring user intervention to be adapted to new skeletons. In this paper we aim at creating a flexible, learned IK solver applicable to a wide variety of human morphologies. We extend a state-of-the-art machine learning IK solver to operate on the well known Skinned Multi-Person Linear model (SMPL). We call our model SMPL-IK, and show that when integrated into real-time 3D software, this extended system opens up opportunities for defining novel AI-assisted animation workflows. For example, pose authoring can be made more flexible with SMPL-IK by allowing users to modify gender and body shape while posing a character. Additionally, when chained with existing pose estimation algorithms, SMPL-IK accelerates posing by allowing users to bootstrap 3D scenes from 2D images while allowing for further editing. Finally, we propose a novel SMPL Shape Inversion mechanism (SMPL-SI) to map arbitrary humanoid characters to the SMPL space, allowing artists to leverage SMPL-IK on custom characters. In addition to qualitative demos showing proposed tools, we present quantitative SMPL-IK baselines on the H36M and AMASS datasets.

Nitpreet Bamra, Vikram Voleti, Alexander Wong et al.

Climate change is increasing the frequency and severity of harmful algal blooms (HABs), which cause significant fish deaths in aquaculture farms. This contributes to ocean pollution and greenhouse gas (GHG) emissions since dead fish are either dumped into the ocean or taken to landfills, which in turn negatively impacts the climate. Currently, the standard method to enumerate harmful algae and other phytoplankton is to manually observe and count them under a microscope. This is a time-consuming, tedious and error-prone process, resulting in compromised management decisions by farmers. Hence, automating this process for quick and accurate HAB monitoring is extremely helpful. However, this requires large and diverse datasets of phytoplankton images, and such datasets are hard to produce quickly. In this work, we explore the feasibility of generating novel high-resolution photorealistic synthetic phytoplankton images, containing multiple species in the same image, given a small dataset of real images. To this end, we employ Generative Adversarial Networks (GANs) to generate synthetic images. We evaluate three different GAN architectures: ProjectedGAN, FastGAN, and StyleGANv2 using standard image quality metrics. We empirically show the generation of high-fidelity synthetic phytoplankton images using a training dataset of only 961 real images. Thus, this work demonstrates the ability of GANs to create large synthetic datasets of phytoplankton from small training datasets, accomplishing a key step towards sustainable systematic monitoring of harmful algal blooms.

Daniel Zhang, Vikram Voleti, Alexander Wong et al.

Creating high-performance generalizable deep neural networks for phytoplankton monitoring requires utilizing large-scale data coming from diverse global water sources. A major challenge to training such networks lies in data privacy, where data collected at different facilities are often restricted from being transferred to a centralized location. A promising approach to overcome this challenge is federated learning, where training is done at site level on local data, and only the model parameters are exchanged over the network to generate a global model. In this study, we explore the feasibility of leveraging federated learning for privacy-preserving training of deep neural networks for phytoplankton classification. More specifically, we simulate two different federated learning frameworks, federated learning (FL) and mutually exclusive FL (ME-FL), and compare their performance to a traditional centralized learning (CL) framework. Experimental results from this study demonstrate the feasibility and potential of federated learning for phytoplankton monitoring.

Mark Boss, Vikram Voleti, Simon Donné et al.

The key-value (KV) cache dominates memory bandwidth and footprint in long-context autoregressive inference. Recent rotation-preconditioned codecs (TurboQuant, PolarQuant) show that a structured random rotation followed by a per-coordinate scalar quantizer matched to an analytically tractable marginal is a near-optimal recipe for KV compression. OCTOPUS advances this paradigm through joint quantization of rotated coordinate triplets. Each triplet's direction is mapped to a square via an octahedral parameterization, and the two resulting coordinates and the triplet norm are Lloyd-Max quantized against implementation-matched marginals. Optimizing the per-triplet squared error gives a strictly non-uniform bit allocation depending only on the total dimensionality of the keys. We find the finite-dimensional quality optimum with sweeps to be constant on every real decoder we test. The codec is data-oblivious, online, and deterministic given a seed. Across text, video, and audio, OCTOPUS matches or beats every prior rotation codec at every reported bit width and metric, with a lead that grows as bits drop for extreme compression. Furthermore, a fused Triton implementation reconstructs keys on the fly without materializing the uncompressed key, so the codec adds no decode-time bandwidth or latency over the existing dequantization. Project Page: https://octopus-quant.github.io/

Vikram Voleti

This dissertation attempts to drive innovation in the field of generative modeling for computer vision, by exploring novel formulations of conditional generative models, and innovative applications in images, 3D animations, and video. Our research focuses on architectures that offer reversible transformations of noise and visual data, and the application of encoder-decoder architectures for generative tasks and 3D content manipulation. In all instances, we incorporate conditional information to enhance the synthesis of visual data, improving the efficiency of the generation process as well as the generated content. We introduce the use of Neural ODEs to model video dynamics using an encoder-decoder architecture, demonstrating their ability to predict future video frames despite being trained solely to reconstruct current frames. Next, we propose a conditional variant of continuous normalizing flows that enables higher-resolution image generation based on lower-resolution input, achieving comparable image quality while reducing parameters and training time. Our next contribution presents a pipeline that takes human images as input, automatically aligns a user-specified 3D character with the pose of the human, and facilitates pose editing based on partial inputs. Next, we derive the relevant mathematical details for denoising diffusion models that use non-isotropic Gaussian processes, and show comparable generation quality. Finally, we devise a novel denoising diffusion framework capable of solving all three video tasks of prediction, generation, and interpolation. We perform ablation studies, and show SOTA results on multiple datasets. Our contributions are published articles at peer-reviewed venues. Overall, our research aims to make a meaningful contribution to the pursuit of more efficient and flexible generative models, with the potential to shape the future of computer vision.

Vikram Voleti, Chun-Han Yao, Mark Boss et al.

We present Stable Video 3D (SV3D) -- a latent video diffusion model for high-resolution, image-to-multi-view generation of orbital videos around a 3D object. Recent work on 3D generation propose techniques to adapt 2D generative models for novel view synthesis (NVS) and 3D optimization. However, these methods have several disadvantages due to either limited views or inconsistent NVS, thereby affecting the performance of 3D object generation. In this work, we propose SV3D that adapts image-to-video diffusion model for novel multi-view synthesis and 3D generation, thereby leveraging the generalization and multi-view consistency of the video models, while further adding explicit camera control for NVS. We also propose improved 3D optimization techniques to use SV3D and its NVS outputs for image-to-3D generation. Extensive experimental results on multiple datasets with 2D and 3D metrics as well as user study demonstrate SV3D's state-of-the-art performance on NVS as well as 3D reconstruction compared to prior works.

Jensen Zhou, Hang Gao, Vikram Voleti et al.

We present Stable Virtual Camera (Seva), a generalist diffusion model that creates novel views of a scene, given any number of input views and target cameras. Existing works struggle to generate either large viewpoint changes or temporally smooth samples, while relying on specific task configurations. Our approach overcomes these limitations through simple model design, optimized training recipe, and flexible sampling strategy that generalize across view synthesis tasks at test time. As a result, our samples maintain high consistency without requiring additional 3D representation-based distillation, thus streamlining view synthesis in the wild. Furthermore, we show that our method can generate high-quality videos lasting up to half a minute with seamless loop closure. Extensive benchmarking demonstrates that Seva outperforms existing methods across different datasets and settings. Project page with code and model: https://stable-virtual-camera.github.io/.

Chun-Han Yao, Yiming Xie, Vikram Voleti et al.

We present Stable Video 4D 2.0 (SV4D 2.0), a multi-view video diffusion model for dynamic 3D asset generation. Compared to its predecessor SV4D, SV4D 2.0 is more robust to occlusions and large motion, generalizes better to real-world videos, and produces higher-quality outputs in terms of detail sharpness and spatio-temporal consistency. We achieve this by introducing key improvements in multiple aspects: 1) network architecture: eliminating the dependency of reference multi-views and designing blending mechanism for 3D and frame attention, 2) data: enhancing quality and quantity of training data, 3) training strategy: adopting progressive 3D-4D training for better generalization, and 4) 4D optimization: handling 3D inconsistency and large motion via 2-stage refinement and progressive frame sampling. Extensive experiments demonstrate significant performance gain by SV4D 2.0 both visually and quantitatively, achieving better detail (-14\% LPIPS) and 4D consistency (-44\% FV4D) in novel-view video synthesis and 4D optimization (-12\% LPIPS and -24\% FV4D) compared to SV4D. Project page: https://sv4d20.github.io.

Andreas Engelhardt, Mark Boss, Vikram Voleti et al.

We present Stable Video Materials 3D (SViM3D), a framework to predict multi-view consistent physically based rendering (PBR) materials, given a single image. Recently, video diffusion models have been successfully used to reconstruct 3D objects from a single image efficiently. However, reflectance is still represented by simple material models or needs to be estimated in additional steps to enable relighting and controlled appearance edits. We extend a latent video diffusion model to output spatially varying PBR parameters and surface normals jointly with each generated view based on explicit camera control. This unique setup allows for relighting and generating a 3D asset using our model as neural prior. We introduce various mechanisms to this pipeline that improve quality in this ill-posed setting. We show state-of-the-art relighting and novel view synthesis performance on multiple object-centric datasets. Our method generalizes to diverse inputs, enabling the generation of relightable 3D assets useful in AR/VR, movies, games and other visual media.

Mallikarjun B. R., Fei Yin, Vikram Voleti et al.

Portrait animation aims to generate photo-realistic videos from a single source image by reenacting the expression and pose from a driving video. While early methods relied on 3D morphable models or feature warping techniques, they often suffered from limited expressivity, temporal inconsistency, and poor generalization to unseen identities or large pose variations. Recent advances using diffusion models have demonstrated improved quality but remain constrained by weak control signals and architectural limitations. In this work, we propose a novel diffusion based framework that leverages masked facial regions specifically the eyes, nose, and mouth from the driving video as strong motion control cues. To enable robust training without appearance leakage, we adopt cross identity supervision. To leverage the strong prior from the pretrained diffusion model, our novel architecture introduces minimal new parameters that converge faster and help in better generalization. We introduce spatial temporal attention mechanisms that allow inter frame and intra frame interactions, effectively capturing subtle motions and reducing temporal artifacts. Our model uses history frames to ensure continuity across segments. At inference, we propose a novel signal fusion strategy that balances motion fidelity with identity preservation. Our approach achieves superior temporal consistency and accurate expression control, enabling high-quality, controllable portrait animation suitable for real-world applications.

Hieu T. Nguyen, Yiwen Chen, Vikram Voleti et al.

We introduce HouseCrafter, a novel approach that can lift a floorplan into a complete large 3D indoor scene (e.g., a house). Our key insight is to adapt a 2D diffusion model, which is trained on web-scale images, to generate consistent multi-view color (RGB) and depth (D) images across different locations of the scene. Specifically, the RGB-D images are generated autoregressively in a batch-wise manner along sampled locations based on the floorplan, where previously generated images are used as condition to the diffusion model to produce images at nearby locations. The global floorplan and attention design in the diffusion model ensures the consistency of the generated images, from which a 3D scene can be reconstructed. Through extensive evaluation on the 3D-Front dataset, we demonstrate that HouseCraft can generate high-quality house-scale 3D scenes. Ablation studies also validate the effectiveness of different design choices. We will release our code and model weights. Project page: https://neu-vi.github.io/houseCrafter/

Benno Krojer, Elinor Poole-Dayan, Vikram Voleti et al.

Text-conditioned image generation models have recently shown immense qualitative success using denoising diffusion processes. However, unlike discriminative vision-and-language models, it is a non-trivial task to subject these diffusion-based generative models to automatic fine-grained quantitative evaluation of high-level phenomena such as compositionality. Towards this goal, we perform two innovations. First, we transform diffusion-based models (in our case, Stable Diffusion) for any image-text matching (ITM) task using a novel method called DiffusionITM. Second, we introduce the Generative-Discriminative Evaluation Benchmark (GDBench) benchmark with 7 complex vision-and-language tasks, bias evaluation and detailed analysis. We find that Stable Diffusion + DiffusionITM is competitive on many tasks and outperforms CLIP on compositional tasks like like CLEVR and Winoground. We further boost its compositional performance with a transfer setup by fine-tuning on MS-COCO while retaining generative capabilities. We also measure the stereotypical bias in diffusion models, and find that Stable Diffusion 2.1 is, for the most part, less biased than Stable Diffusion 1.5. Overall, our results point in an exciting direction bringing discriminative and generative model evaluation closer. We will release code and benchmark setup soon.

David Kanaa, Vikram Voleti, Samira Ebrahimi Kahou et al.

Despite having been studied to a great extent, the task of conditional generation of sequences of frames, or videos, remains extremely challenging. It is a common belief that a key step towards solving this task resides in modelling accurately both spatial and temporal information in video signals. A promising direction to do so has been to learn latent variable models that predict the future in latent space and project back to pixels, as suggested in recent literature. Following this line of work and building on top of a family of models introduced in prior work, Neural ODE, we investigate an approach that models time-continuous dynamics over a continuous latent space with a differential equation with respect to time. The intuition behind this approach is that these trajectories in latent space could then be extrapolated to generate video frames beyond the time steps for which the model is trained. We show that our approach yields promising results in the task of future frame prediction on the Moving MNIST dataset with 1 and 2 digits.

Ju An Park, Vikram Voleti, Kathryn E. Thomas et al.

Warming oceans due to climate change are leading to increased numbers of ectoparasitic copepods, also known as sea lice, which can cause significant ecological loss to wild salmon populations and major economic loss to aquaculture sites. The main transport mechanism driving the spread of sea lice populations are near-surface ocean currents. Present strategies to estimate the distribution of sea lice larvae are computationally complex and limit full-scale analysis. Motivated to address this challenge, we propose SALT: Sea lice Adaptive Lattice Tracking approach for efficient estimation of sea lice dispersion and distribution in space and time. Specifically, an adaptive spatial mesh is generated by merging nodes in the lattice graph of the Ocean Model based on local ocean properties, thus enabling highly efficient graph representation. SALT demonstrates improved efficiency while maintaining consistent results with the standard method, using near-surface current data for Hardangerfjord, Norway. The proposed SALT technique shows promise for enhancing proactive aquaculture management through predictive modelling of sea lice infestation pressure maps in a changing climate.

Vikram Voleti, Chris Finlay, Adam Oberman et al.

Recent work has shown that Neural Ordinary Differential Equations (ODEs) can serve as generative models of images using the perspective of Continuous Normalizing Flows (CNFs). Such models offer exact likelihood calculation, and invertible generation/density estimation. In this work we introduce a Multi-Resolution variant of such models (MRCNF), by characterizing the conditional distribution over the additional information required to generate a fine image that is consistent with the coarse image. We introduce a transformation between resolutions that allows for no change in the log likelihood. We show that this approach yields comparable likelihood values for various image datasets, with improved performance at higher resolutions, with fewer parameters, using only 1 GPU. Further, we examine the out-of-distribution properties of (Multi-Resolution) Continuous Normalizing Flows, and find that they are similar to those of other likelihood-based generative models.

Tiago Salvador, Vikram Voleti, Alexander Iannantuono et al.

Distribution shift is an important concern in deep image classification, produced either by corruption of the source images, or a complete change, with the solution involving domain adaptation. While the primary goal is to improve accuracy under distribution shift, an important secondary goal is uncertainty estimation: evaluating the probability that the prediction of a model is correct. While improving accuracy is hard, uncertainty estimation turns out to be frustratingly easy. Prior works have appended uncertainty estimation into the model and training paradigm in various ways. Instead, we show that we can estimate uncertainty by simply exposing the original model to corrupted images, and performing simple statistical calibration on the image outputs. Our frustratingly easy methods demonstrate superior performance on a wide range of distribution shifts as well as on unsupervised domain adaptation tasks, measured through extensive experimentation.

Tiago Salvador, Stephanie Cairns, Vikram Voleti et al.

Despite being widely used, face recognition models suffer from bias: the probability of a false positive (incorrect face match) strongly depends on sensitive attributes such as the ethnicity of the face. As a result, these models can disproportionately and negatively impact minority groups, particularly when used by law enforcement. The majority of bias reduction methods have several drawbacks: they use an end-to-end retraining approach, may not be feasible due to privacy issues, and often reduce accuracy. An alternative approach is post-processing methods that build fairer decision classifiers using the features of pre-trained models, thus avoiding the cost of retraining. However, they still have drawbacks: they reduce accuracy (AGENDA, PASS, FTC), or require retuning for different false positive rates (FSN). In this work, we introduce the Fairness Calibration (FairCal) method, a post-training approach that simultaneously: (i) increases model accuracy (improving the state-of-the-art), (ii) produces fairly-calibrated probabilities, (iii) significantly reduces the gap in the false positive rates, (iv) does not require knowledge of the sensitive attribute, and (v) does not require retraining, training an additional model, or retuning. We apply it to the task of Face Verification, and obtain state-of-the-art results with all the above advantages.

Krishna Murthy Jatavallabhula, Miles Macklin, Florian Golemo et al.

We consider the problem of estimating an object's physical properties such as mass, friction, and elasticity directly from video sequences. Such a system identification problem is fundamentally ill-posed due to the loss of information during image formation. Current solutions require precise 3D labels which are labor-intensive to gather, and infeasible to create for many systems such as deformable solids or cloth. We present gradSim, a framework that overcomes the dependence on 3D supervision by leveraging differentiable multiphysics simulation and differentiable rendering to jointly model the evolution of scene dynamics and image formation. This novel combination enables backpropagation from pixels in a video sequence through to the underlying physical attributes that generated them. Moreover, our unified computation graph -- spanning from the dynamics and through the rendering process -- enables learning in challenging visuomotor control tasks, without relying on state-based (3D) supervision, while obtaining performance competitive to or better than techniques that rely on precise 3D labels.

Xavier Bouthillier, Pierre Delaunay, Mirko Bronzi et al.

Strong empirical evidence that one machine-learning algorithm A outperforms another one B ideally calls for multiple trials optimizing the learning pipeline over sources of variation such as data sampling, data augmentation, parameter initialization, and hyperparameters choices. This is prohibitively expensive, and corners are cut to reach conclusions. We model the whole benchmarking process, revealing that variance due to data sampling, parameter initialization and hyperparameter choice impact markedly the results. We analyze the predominant comparison methods used today in the light of this variance. We show a counter-intuitive result that adding more sources of variation to an imperfect estimator approaches better the ideal estimator at a 51 times reduction in compute cost. Building on these results, we study the error rate of detecting improvements, on five different deep-learning tasks/architectures. This study leads us to propose recommendations for performance comparisons.

Sarthak Mittal, Alex Lamb, Anirudh Goyal et al.

Robust perception relies on both bottom-up and top-down signals. Bottom-up signals consist of what's directly observed through sensation. Top-down signals consist of beliefs and expectations based on past experience and short-term memory, such as how the phrase `peanut butter and~...' will be completed. The optimal combination of bottom-up and top-down information remains an open question, but the manner of combination must be dynamic and both context and task dependent. To effectively utilize the wealth of potential top-down information available, and to prevent the cacophony of intermixed signals in a bidirectional architecture, mechanisms are needed to restrict information flow. We explore deep recurrent neural net architectures in which bottom-up and top-down signals are dynamically combined using attention. Modularity of the architecture further restricts the sharing and communication of information. Together, attention and modularity direct information flow, which leads to reliable performance improvements in perceptual and language tasks, and in particular improves robustness to distractions and noisy data. We demonstrate on a variety of benchmarks in language modeling, sequential image classification, video prediction and reinforcement learning that the \emph{bidirectional} information flow can improve results over strong baselines.

Vincent Michalski, Vikram Voleti, Samira Ebrahimi Kahou et al.

Batch normalization has been widely used to improve optimization in deep neural networks. While the uncertainty in batch statistics can act as a regularizer, using these dataset statistics specific to the training set impairs generalization in certain tasks. Recently, alternative methods for normalizing feature activations in neural networks have been proposed. Among them, group normalization has been shown to yield similar, in some domains even superior performance to batch normalization. All these methods utilize a learned affine transformation after the normalization operation to increase representational power. Methods used in conditional computation define the parameters of these transformations as learnable functions of conditioning information. In this work, we study whether and where the conditional formulation of group normalization can improve generalization compared to conditional batch normalization. We evaluate performances on the tasks of visual question answering, few-shot learning, and conditional image generation.