Visual understanding of geometric structures with complex spatial relationships is a fundamental component of human intelligence. As children, we learn how to reason about structure not only from observation, but also by interacting with the world around us -- by taking things apart and putting them back together again. The ability to reason about structure and compositionality allows us to not only build things, but also understand and reverse-engineer complex systems. In order to advance research in interactive reasoning for part-based geometric understanding, we propose a challenging new assembly problem using LEGO bricks that we call Break and Make. In this problem an agent is given a LEGO model and attempts to understand its structure by interactively inspecting and disassembling it. After this inspection period, the agent must then prove its understanding by rebuilding the model from scratch using low-level action primitives. In order to facilitate research on this problem we have built LTRON, a fully interactive 3D simulator that allows learning agents to assemble, disassemble and manipulate LEGO models. We pair this simulator with a new dataset of fan-made LEGO creations that have been uploaded to the internet in order to provide complex scenes containing over a thousand unique brick shapes. We take a first step towards solving this problem using sequence-to-sequence models that provide guidance for how to make progress on this challenging problem. Our simulator and data are available at github.com/aaronwalsman/ltron. Additional training code and PyTorch examples are available at github.com/aaronwalsman/ltron-torch-eccv22.
Massive datasets and high-capacity models have driven many recent advancements in computer vision and natural language understanding. This work presents a platform to enable similar success stories in Embodied AI. We propose ProcTHOR, a framework for procedural generation of Embodied AI environments. ProcTHOR enables us to sample arbitrarily large datasets of diverse, interactive, customizable, and performant virtual environments to train and evaluate embodied agents across navigation, interaction, and manipulation tasks. We demonstrate the power and potential of ProcTHOR via a sample of 10,000 generated houses and a simple neural model. Models trained using only RGB images on ProcTHOR, with no explicit mapping and no human task supervision produce state-of-the-art results across 6 embodied AI benchmarks for navigation, rearrangement, and arm manipulation, including the presently running Habitat 2022, AI2-THOR Rearrangement 2022, and RoboTHOR challenges. We also demonstrate strong 0-shot results on these benchmarks, via pre-training on ProcTHOR with no fine-tuning on the downstream benchmark, often beating previous state-of-the-art systems that access the downstream training data.
Learned representations are a central component in modern ML systems, serving a multitude of downstream tasks. When training such representations, it is often the case that computational and statistical constraints for each downstream task are unknown. In this context rigid, fixed capacity representations can be either over or under-accommodating to the task at hand. This leads us to ask: can we design a flexible representation that can adapt to multiple downstream tasks with varying computational resources? Our main contribution is Matryoshka Representation Learning (MRL) which encodes information at different granularities and allows a single embedding to adapt to the computational constraints of downstream tasks. MRL minimally modifies existing representation learning pipelines and imposes no additional cost during inference and deployment. MRL learns coarse-to-fine representations that are at least as accurate and rich as independently trained low-dimensional representations. The flexibility within the learned Matryoshka Representations offer: (a) up to 14x smaller embedding size for ImageNet-1K classification at the same level of accuracy; (b) up to 14x real-world speed-ups for large-scale retrieval on ImageNet-1K and 4K; and (c) up to 2% accuracy improvements for long-tail few-shot classification, all while being as robust as the original representations. Finally, we show that MRL extends seamlessly to web-scale datasets (ImageNet, JFT) across various modalities -- vision (ViT, ResNet), vision + language (ALIGN) and language (BERT). MRL code and pretrained models are open-sourced at https://github.com/RAIVNLab/MRL.
Object manipulation is a critical skill required for Embodied AI agents interacting with the world around them. Training agents to manipulate objects, poses many challenges. These include occlusion of the target object by the agent's arm, noisy object detection and localization, and the target frequently going out of view as the agent moves around in the scene. We propose Manipulation via Visual Object Location Estimation (m-VOLE), an approach that explores the environment in search for target objects, computes their 3D coordinates once they are located, and then continues to estimate their 3D locations even when the objects are not visible, thus robustly aiding the task of manipulating these objects throughout the episode. Our evaluations show a massive 3x improvement in success rate over a model that has access to the same sensory suite but is trained without the object location estimator, and our analysis shows that our agent is robust to noise in depth perception and agent localization. Importantly, our proposed approach relaxes several assumptions about idealized localization and perception that are commonly employed by recent works in embodied AI -- an important step towards training agents for object manipulation in the real world.
The conventional recipe for maximizing model accuracy is to (1) train multiple models with various hyperparameters and (2) pick the individual model which performs best on a held-out validation set, discarding the remainder. In this paper, we revisit the second step of this procedure in the context of fine-tuning large pre-trained models, where fine-tuned models often appear to lie in a single low error basin. We show that averaging the weights of multiple models fine-tuned with different hyperparameter configurations often improves accuracy and robustness. Unlike a conventional ensemble, we may average many models without incurring any additional inference or memory costs -- we call the results "model soups." When fine-tuning large pre-trained models such as CLIP, ALIGN, and a ViT-G pre-trained on JFT, our soup recipe provides significant improvements over the best model in a hyperparameter sweep on ImageNet. As a highlight, the resulting ViT-G model attains 90.94% top-1 accuracy on ImageNet, a new state of the art. Furthermore, we show that the model soup approach extends to multiple image classification and natural language processing tasks, improves out-of-distribution performance, and improves zero-shot performance on new downstream tasks. Finally, we analytically relate the performance similarity of weight-averaging and logit-ensembling to flatness of the loss and confidence of the predictions, and validate this relation empirically.
As humans, we navigate the world through all our senses, using perceptual input from each one to correct the others. We introduce MERLOT Reserve, a model that represents videos jointly over time -- through a new training objective that learns from audio, subtitles, and video frames. Given a video, we replace snippets of text and audio with a MASK token; the model learns by choosing the correct masked-out snippet. Our objective learns faster than alternatives, and performs well at scale: we pretrain on 20 million YouTube videos. Empirical results show that MERLOT Reserve learns strong representations about videos through all constituent modalities. When finetuned, it sets a new state-of-the-art on both VCR and TVQA, outperforming prior work by 5% and 7% respectively. Ablations show that both tasks benefit from audio pretraining -- even VCR, a QA task centered around images (without sound). Moreover, our objective enables out-of-the-box prediction, revealing strong multimodal commonsense understanding. In a fully zero-shot setting, our model obtains competitive results on four video understanding tasks, even outperforming supervised approaches on the recently proposed Situated Reasoning (STAR) benchmark. We analyze why incorporating audio leads to better vision-language representations, suggesting significant opportunities for future research. We conclude by discussing ethical and societal implications of multimodal pretraining.
The last few years have witnessed substantial progress in the field of embodied AI where artificial agents, mirroring biological counterparts, are now able to learn from interaction to accomplish complex tasks. Despite this success, biological organisms still hold one large advantage over these simulated agents: adaptation. While both living and simulated agents make decisions to achieve goals (strategy), biological organisms have evolved to understand their environment (sensing) and respond to it (physiology). The net gain of these factors depends on the environment, and organisms have adapted accordingly. For example, in a low vision aquatic environment some fish have evolved specific neurons which offer a predictable, but incredibly rapid, strategy to escape from predators. Mammals have lost these reactive systems, but they have a much larger fields of view and brain circuitry capable of understanding many future possibilities. While traditional embodied agents manipulate an environment to best achieve a goal, we argue for an introspective agent, which considers its own abilities in the context of its environment. We show that different environments yield vastly different optimal designs, and increasing long-term planning is often far less beneficial than other improvements, such as increased physical ability. We present these findings to broaden the definition of improvement in embodied AI passed increasingly complex models. Just as in nature, we hope to reframe strategy as one tool, among many, to succeed in an environment. Code is available at: https://github.com/sarahpratt/introspective.
In visual retrieval systems, updating the embedding model requires recomputing features for every piece of data. This expensive process is referred to as backfilling. Recently, the idea of backward compatible training (BCT) was proposed. To avoid the cost of backfilling, BCT modifies training of the new model to make its representations compatible with those of the old model. However, BCT can significantly hinder the performance of the new model. In this work, we propose a new learning paradigm for representation learning: forward compatible training (FCT). In FCT, when the old model is trained, we also prepare for a future unknown version of the model. We propose learning side-information, an auxiliary feature for each sample which facilitates future updates of the model. To develop a powerful and flexible framework for model compatibility, we combine side-information with a forward transformation from old to new embeddings. Training of the new model is not modified, hence, its accuracy is not degraded. We demonstrate significant retrieval accuracy improvement compared to BCT for various datasets: ImageNet-1k (+18.1%), Places-365 (+5.4%), and VGG-Face2 (+8.3%). FCT obtains model compatibility when the new and old models are trained across different datasets, losses, and architectures.
Communicating with humans is challenging for AIs because it requires a shared understanding of the world, complex semantics (e.g., metaphors or analogies), and at times multi-modal gestures (e.g., pointing with a finger, or an arrow in a diagram). We investigate these challenges in the context of Iconary, a collaborative game of drawing and guessing based on Pictionary, that poses a novel challenge for the research community. In Iconary, a Guesser tries to identify a phrase that a Drawer is drawing by composing icons, and the Drawer iteratively revises the drawing to help the Guesser in response. This back-and-forth often uses canonical scenes, visual metaphor, or icon compositions to express challenging words, making it an ideal test for mixing language and visual/symbolic communication in AI. We propose models to play Iconary and train them on over 55,000 games between human players. Our models are skillful players and are able to employ world knowledge in language models to play with words unseen during training. Elite human players outperform our models, particularly at the drawing task, leaving an important gap for future research to address. We release our dataset, code, and evaluation setup as a challenge to the community at http://www.github.com/allenai/iconary.
When deploying deep learning models to a device, it is traditionally assumed that available computational resources (compute, memory, and power) remain static. However, real-world computing systems do not always provide stable resource guarantees. Computational resources need to be conserved when load from other processes is high or battery power is low. Inspired by recent works on neural network subspaces, we propose a method for training a "compressible subspace" of neural networks that contains a fine-grained spectrum of models that range from highly efficient to highly accurate. Our models require no retraining, thus our subspace of models can be deployed entirely on-device to allow adaptive network compression at inference time. We present results for achieving arbitrarily fine-grained accuracy-efficiency trade-offs at inference time for structured and unstructured sparsity. We achieve accuracies on-par with standard models when testing our uncompressed models, and maintain high accuracy for sparsity rates above 90% when testing our compressed models. We also demonstrate that our algorithm extends to quantization at variable bit widths, achieving accuracy on par with individually trained networks.