Jonathan Huang

Jae Joong Lee, Bosheng Li, Sara Beery et al. · mit

We introduce Tree D-fusion, featuring the first collection of 600,000 environmentally aware, 3D simulation-ready tree models generated through Diffusion priors. Each reconstructed 3D tree model corresponds to an image from Google's Auto Arborist Dataset, comprising street view images and associated genus labels of trees across North America. Our method distills the scores of two tree-adapted diffusion models by utilizing text prompts to specify a tree genus, thus facilitating shape reconstruction. This process involves reconstructing a 3D tree envelope filled with point markers, which are subsequently utilized to estimate the tree's branching structure using the space colonization algorithm conditioned on a specified genus.

Xiuye Gu, Yin Cui, Jonathan Huang et al.

Observing the close relationship among panoptic, semantic and instance segmentation tasks, we propose to train a universal multi-dataset multi-task segmentation model: DaTaSeg.We use a shared representation (mask proposals with class predictions) for all tasks. To tackle task discrepancy, we adopt different merge operations and post-processing for different tasks. We also leverage weak-supervision, allowing our segmentation model to benefit from cheaper bounding box annotations. To share knowledge across datasets, we use text embeddings from the same semantic embedding space as classifiers and share all network parameters among datasets. We train DaTaSeg on ADE semantic, COCO panoptic, and Objects365 detection datasets. DaTaSeg improves performance on all datasets, especially small-scale datasets, achieving 54.0 mIoU on ADE semantic and 53.5 PQ on COCO panoptic. DaTaSeg also enables weakly-supervised knowledge transfer on ADE panoptic and Objects365 instance segmentation. Experiments show DaTaSeg scales with the number of training datasets and enables open-vocabulary segmentation through direct transfer. In addition, we annotate an Objects365 instance segmentation set of 1,000 images and will release it as a public benchmark.

Nikolai Warner, Meera Hahn, Jonathan Huang et al.

Segmentation localizes objects in an image on a fine-grained per-pixel scale. Segmentation benefits by humans-in-the-loop to provide additional input of objects to segment using a combination of foreground or background clicks. Tasks include photoediting or novel dataset annotation, where human annotators leverage an existing segmentation model instead of drawing raw pixel level annotations. We propose a new segmentation process, Text + Click segmentation, where a model takes as input an image, a text phrase describing a class to segment, and a single foreground click specifying the instance to segment. Compared to previous approaches, we leverage open-vocabulary image-text models to support a wide-range of text prompts. Conditioning segmentations on text prompts improves the accuracy of segmentations on novel or unseen classes. We demonstrate that the combination of a single user-specified foreground click and a text prompt allows a model to better disambiguate overlapping or co-occurring semantic categories, such as "tie", "suit", and "person". We study these results across common segmentation datasets such as refCOCO, COCO, VOC, and OpenImages. Source code available here.

Shreyank N Gowda, Anurag Arnab, Jonathan Huang

In this paper, we address the challenges posed by the substantial training time and memory consumption associated with video transformers, focusing on the ViViT (Video Vision Transformer) model, in particular the Factorised Encoder version, as our baseline for action recognition tasks. The factorised encoder variant follows the late-fusion approach that is adopted by many state of the art approaches. Despite standing out for its favorable speed/accuracy tradeoffs among the different variants of ViViT, its considerable training time and memory requirements still pose a significant barrier to entry. Our method is designed to lower this barrier and is based on the idea of freezing the spatial transformer during training. This leads to a low accuracy model if naively done. But we show that by (1) appropriately initializing the temporal transformer (a module responsible for processing temporal information) (2) introducing a compact adapter model connecting frozen spatial representations ((a module that selectively focuses on regions of the input image) to the temporal transformer, we can enjoy the benefits of freezing the spatial transformer without sacrificing accuracy. Through extensive experimentation over 6 benchmarks, we demonstrate that our proposed training strategy significantly reduces training costs (by $\sim 50\%$) and memory consumption while maintaining or slightly improving performance by up to 1.79\% compared to the baseline model. Our approach additionally unlocks the capability to utilize larger image transformer models as our spatial transformer and access more frames with the same memory consumption.

Dan Kondratyuk, Lijun Yu, Xiuye Gu et al. · cmu, deepmind

We present VideoPoet, a language model capable of synthesizing high-quality video, with matching audio, from a large variety of conditioning signals. VideoPoet employs a decoder-only transformer architecture that processes multimodal inputs -- including images, videos, text, and audio. The training protocol follows that of Large Language Models (LLMs), consisting of two stages: pretraining and task-specific adaptation. During pretraining, VideoPoet incorporates a mixture of multimodal generative objectives within an autoregressive Transformer framework. The pretrained LLM serves as a foundation that can be adapted for a range of video generation tasks. We present empirical results demonstrating the model's state-of-the-art capabilities in zero-shot video generation, specifically highlighting VideoPoet's ability to generate high-fidelity motions. Project page: http://sites.research.google/videopoet/

Junhua Mao, Jonathan Huang, Alexander Toshev et al.

We propose a method that can generate an unambiguous description (known as a referring expression) of a specific object or region in an image, and which can also comprehend or interpret such an expression to infer which object is being described. We show that our method outperforms previous methods that generate descriptions of objects without taking into account other potentially ambiguous objects in the scene. Our model is inspired by recent successes of deep learning methods for image captioning, but while image captioning is difficult to evaluate, our task allows for easy objective evaluation. We also present a new large-scale dataset for referring expressions, based on MS-COCO. We have released the dataset and a toolbox for visualization and evaluation, see https://github.com/mjhucla/Google_Refexp_toolbox

Sihyun Yu, Meera Hahn, Dan Kondratyuk et al.

Diffusion models are successful for synthesizing high-quality videos but are limited to generating short clips (e.g., 2-10 seconds). Synthesizing sustained footage (e.g. over minutes) still remains an open research question. In this paper, we propose MALT Diffusion (using Memory-Augmented Latent Transformers), a new diffusion model specialized for long video generation. MALT Diffusion (or just MALT) handles long videos by subdividing them into short segments and doing segment-level autoregressive generation. To achieve this, we first propose recurrent attention layers that encode multiple segments into a compact memory latent vector; by maintaining this memory vector over time, MALT is able to condition on it and continuously generate new footage based on a long temporal context. We also present several training techniques that enable the model to generate frames over a long horizon with consistent quality and minimal degradation. We validate the effectiveness of MALT through experiments on long video benchmarks. We first perform extensive analysis of MALT in long-contextual understanding capability and stability using popular long video benchmarks. For example, MALT achieves an FVD score of 220.4 on 128-frame video generation on UCF-101, outperforming the previous state-of-the-art of 648.4. Finally, we explore MALT's capabilities in a text-to-video generation setting and show that it can produce long videos compared with recent techniques for long text-to-video generation.

Evandros Kaklamanos, Kristjana Kristinsdottir, Jonathan Huang et al.

Endoscopic procedures such as esophagogastroduodenoscopy (EGD) and colonoscopy play a critical role in diagnosing and managing gastrointestinal (GI) disorders. However, the documentation burden associated with these procedures place significant strain on gastroenterologists, contributing to inefficiencies in clinical workflows and physician burnout. To address this challenge, we propose a novel automated report generation model that leverages a transformer-based vision encoder and text decoder within a two-stage training framework. In the first stage, both components are pre-trained on image/text caption pairs to capture generalized vision-language features, followed by fine-tuning on images/report pairs to generate clinically meaningful findings. Our approach not only streamlines the documentation process but also holds promise for reducing physician workload and improving patient care.

Nikolai Warner, Jack Kolb, Meera Hahn et al.

In this paper we focus on inserting a given human (specifically, a single image of a person) into a novel scene. Our method, which builds on top of Stable Diffusion, yields natural looking images while being highly controllable with text and pose. To accomplish this we need to train on pairs of images, the first a reference image with the person, the second a "target image" showing the same person (with a different pose and possibly in a different background). Additionally we require a text caption describing the new pose relative to that in the reference image. In this paper we present a novel dataset following this criteria, which we create using pairs of frames from human-centric and action-rich videos and employing a multimodal LLM to automatically summarize the difference in human pose for the text captions. We demonstrate that identity preservation is a more challenging task in scenes "in-the-wild", and especially scenes where there is an interaction between persons and objects. Combining the weak supervision from noisy captions, with robust 2D pose improves the quality of person-object interactions.

Xinyue Hao, Gen Li, Shreyank N Gowda et al.

Video understanding has made huge strides in recent years, relying largely on the power of transformers. As this architecture is notoriously expensive and video data is highly redundant, research into improving efficiency has become particularly relevant. Some creative solutions include token selection and merging. While most methods succeed in reducing the cost of the model and maintaining accuracy, an interesting pattern arises: most methods do not outperform the baseline of randomly discarding tokens. In this paper we take a closer look at this phenomenon and observe 5 principles of the nature of visual tokens. For example, we observe that the value of tokens follows a clear Pareto-distribution where most tokens have remarkably low value, and just a few carry most of the perceptual information. We build on these and further insights to propose a lightweight video model, LITE, that can select a small number of tokens effectively, outperforming state-of-the-art and existing baselines across datasets (Kinetics-400 and Something-Something-V2) in the challenging trade-off of computation (GFLOPs) vs accuracy. Experiments also show that LITE generalizes across datasets and even other tasks without the need for retraining.

Seulki Park, Youren Zhang, Stella X. Yu et al.

Hierarchical classification predicts labels across multiple levels of a taxonomy, e.g., from coarse-level 'Bird' to mid-level 'Hummingbird' to fine-level 'Green hermit', allowing flexible recognition under varying visual conditions. It is commonly framed as multiple single-level tasks, but each level may rely on different visual cues: Distinguishing 'Bird' from 'Plant' relies on global features like feathers or leaves, while separating 'Anna's hummingbird' from 'Green hermit' requires local details such as head coloration. Prior methods improve accuracy using external semantic supervision, but such statistical learning criteria fail to ensure consistent visual grounding at test time, resulting in incorrect hierarchical classification. We propose, for the first time, to enforce internal visual consistency by aligning fine-to-coarse predictions through intra-image segmentation. Our method outperforms zero-shot CLIP and state-of-the-art baselines on hierarchical classification benchmarks, achieving both higher accuracy and more consistent predictions. It also improves internal image segmentation without requiring pixel-level annotations.

Vasudha Kowtha, Miquel Espi Marques, Jonathan Huang et al.

This work investigates pretrained audio representations for few shot Sound Event Detection. We specifically address the task of few shot detection of novel acoustic sequences, or sound events with semantically meaningful temporal structure, without assuming access to non-target audio. We develop procedures for pretraining suitable representations, and methods which transfer them to our few shot learning scenario. Our experiments evaluate the general purpose utility of our pretrained representations on AudioSet, and the utility of proposed few shot methods via tasks constructed from real-world acoustic sequences. Our pretrained embeddings are suitable to the proposed task, and enable multiple aspects of our few shot framework.

Jack Valmadre, Alex Bewley, Jonathan Huang et al.

This paper introduces temporally local metrics for Multi-Object Tracking. These metrics are obtained by restricting existing metrics based on track matching to a finite temporal horizon, and provide new insight into the ability of trackers to maintain identity over time. Moreover, the horizon parameter offers a novel, meaningful mechanism by which to define the relative importance of detection and association, a common dilemma in applications where imperfect association is tolerable. It is shown that the historical Average Tracking Accuracy (ATA) metric exhibits superior sensitivity to association, enabling its proposed local variant, ALTA, to capture a wide range of characteristics. In particular, ALTA is better equipped to identify advances in association independent of detection. The paper further presents an error decomposition for ATA that reveals the impact of four distinct error types and is equally applicable to ALTA. The diagnostic capabilities of ALTA are demonstrated on the MOT 2017 and Waymo Open Dataset benchmarks.

Vighnesh Birodkar, Zhichao Lu, Siyang Li et al.

Instance segmentation models today are very accurate when trained on large annotated datasets, but collecting mask annotations at scale is prohibitively expensive. We address the partially supervised instance segmentation problem in which one can train on (significantly cheaper) bounding boxes for all categories but use masks only for a subset of categories. In this work, we focus on a popular family of models which apply differentiable cropping to a feature map and predict a mask based on the resulting crop. Under this family, we study Mask R-CNN and discover that instead of its default strategy of training the mask-head with a combination of proposals and groundtruth boxes, training the mask-head with only groundtruth boxes dramatically improves its performance on novel classes. This training strategy also allows us to take advantage of alternative mask-head architectures, which we exploit by replacing the typical mask-head of 2-4 layers with significantly deeper off-the-shelf architectures (e.g. ResNet, Hourglass models). While many of these architectures perform similarly when trained in fully supervised mode, our main finding is that they can generalize to novel classes in dramatically different ways. We call this ability of mask-heads to generalize to unseen classes the strong mask generalization effect and show that without any specialty modules or losses, we can achieve state-of-the-art results in the partially supervised COCO instance segmentation benchmark. Finally, we demonstrate that our effect is general, holding across underlying detection methodologies (including anchor-based, anchor-free or no detector at all) and across different backbone networks. Code and pre-trained models are available at https://git.io/deepmac.

Yinxiao Li, Zhichao Lu, Xuehan Xiong et al.

In recent years, many works in the video action recognition literature have shown that two stream models (combining spatial and temporal input streams) are necessary for achieving state of the art performance. In this paper we show the benefits of including yet another stream based on human pose estimated from each frame -- specifically by rendering pose on input RGB frames. At first blush, this additional stream may seem redundant given that human pose is fully determined by RGB pixel values -- however we show (perhaps surprisingly) that this simple and flexible addition can provide complementary gains. Using this insight, we then propose a new model, which we dub PERF-Net (short for Pose Empowered RGB-Flow Net), which combines this new pose stream with the standard RGB and flow based input streams via distillation techniques and show that our model outperforms the state-of-the-art by a large margin in a number of human action recognition datasets while not requiring flow or pose to be explicitly computed at inference time. The proposed pose stream is also part of the winner solution of the ActivityNet Kinetics Challenge 2020.

Wenda Chen, Jonathan Huang, Mark Hasegawa-Johnson

This paper focuses on wake on intent (WOI) techniques for platforms with limited compute and memory. Our approach of utterance-level intent classification is based on a sequence of keywords in the utterance instead of a single fixed key phrase. The keyword sequence is transformed into four types of input features, namely acoustics, phones, word2vec and speech2vec for individual intent learning and then fused decision making. If a wake intent is detected, it will trigger the power-costly ASR afterwards. The system is trained and tested on a newly collected internal dataset in Intel called AMIE, which will be reported in this paper for the first time. It is demonstrated that our novel technique with the representation of the key-phrases successfully achieved a noise robust intent classification in different domains including in-car human-machine communications. The wake on intent system will be low-power and low-complexity, which makes it suitable for always on operations in real life hardware-based applications.

Munir Georges, Jonathan Huang, Tobias Bocklet

Deep neural networks (DNN) have recently been widely used in speaker recognition systems, achieving state-of-the-art performance on various benchmarks. The x-vector architecture is especially popular in this research community, due to its excellent performance and manageable computational complexity. In this paper, we present the lrx-vector system, which is the low-rank factorized version of the x-vector embedding network. The primary objective of this topology is to further reduce the memory requirement of the speaker recognition system. We discuss the deployment of knowledge distillation for training the lrx-vector system and compare against low-rank factorization with SVD. On the VOiCES 2019 far-field corpus we were able to reduce the weights by 28% compared to the full-rank x-vector system while keeping the recognition rate constant (1.83% EER).

Zhichao Lu, Vivek Rathod, Ronny Votel et al.

Traditionally multi-object tracking and object detection are performed using separate systems with most prior works focusing exclusively on one of these aspects over the other. Tracking systems clearly benefit from having access to accurate detections, however and there is ample evidence in literature that detectors can benefit from tracking which, for example, can help to smooth predictions over time. In this paper we focus on the tracking-by-detection paradigm for autonomous driving where both tasks are mission critical. We propose a conceptually simple and efficient joint model of detection and tracking, called RetinaTrack, which modifies the popular single stage RetinaNet approach such that it is amenable to instance-level embedding training. We show, via evaluations on the Waymo Open Dataset, that we outperform a recent state of the art tracking algorithm while requiring significantly less computation. We believe that our simple yet effective approach can serve as a strong baseline for future work in this area.

Shachi H Kumar, Eda Okur, Saurav Sahay et al.

We are witnessing a confluence of vision, speech and dialog system technologies that are enabling the IVAs to learn audio-visual groundings of utterances and have conversations with users about the objects, activities and events surrounding them. Recent progress in visual grounding techniques and Audio Understanding are enabling machines to understand shared semantic concepts and listen to the various sensory events in the environment. With audio and visual grounding methods, end-to-end multimodal SDS are trained to meaningfully communicate with us in natural language about the real dynamic audio-visual sensory world around us. In this work, we explore the role of `topics' as the context of the conversation along with multimodal attention into such an end-to-end audio-visual scene-aware dialog system architecture. We also incorporate an end-to-end audio classification ConvNet, AclNet, into our models. We develop and test our approaches on the Audio Visual Scene-Aware Dialog (AVSD) dataset released as a part of the DSTC7. We present the analysis of our experiments and show that some of our model variations outperform the baseline system released for AVSD.

Shachi H Kumar, Eda Okur, Saurav Sahay et al.

With the recent advancements in Artificial Intelligence (AI), Intelligent Virtual Assistants (IVA) such as Alexa, Google Home, etc., have become a ubiquitous part of many homes. Currently, such IVAs are mostly audio-based, but going forward, we are witnessing a confluence of vision, speech and dialog system technologies that are enabling the IVAs to learn audio-visual groundings of utterances. This will enable agents to have conversations with users about the objects, activities and events surrounding them. In this work, we present three main architectural explorations for the Audio Visual Scene-Aware Dialog (AVSD): 1) investigating `topics' of the dialog as an important contextual feature for the conversation, 2) exploring several multimodal attention mechanisms during response generation, 3) incorporating an end-to-end audio classification ConvNet, AclNet, into our architecture. We discuss detailed analysis of the experimental results and show that our model variations outperform the baseline system presented for the AVSD task.

Sara Beery, Guanhang Wu, Vivek Rathod et al.

In static monitoring cameras, useful contextual information can stretch far beyond the few seconds typical video understanding models might see: subjects may exhibit similar behavior over multiple days, and background objects remain static. Due to power and storage constraints, sampling frequencies are low, often no faster than one frame per second, and sometimes are irregular due to the use of a motion trigger. In order to perform well in this setting, models must be robust to irregular sampling rates. In this paper we propose a method that leverages temporal context from the unlabeled frames of a novel camera to improve performance at that camera. Specifically, we propose an attention-based approach that allows our model, Context R-CNN, to index into a long term memory bank constructed on a per-camera basis and aggregate contextual features from other frames to boost object detection performance on the current frame. We apply Context R-CNN to two settings: (1) species detection using camera traps, and (2) vehicle detection in traffic cameras, showing in both settings that Context R-CNN leads to performance gains over strong baselines. Moreover, we show that increasing the contextual time horizon leads to improved results. When applied to camera trap data from the Snapshot Serengeti dataset, Context R-CNN with context from up to a month of images outperforms a single-frame baseline by 17.9% mAP, and outperforms S3D (a 3d convolution based baseline) by 11.2% mAP.

Jingchi Zhang, Jonathan Huang, Michael Deisher et al.

Recently, deep neural networks (DNN) have been widely used in speaker recognition area. In order to achieve fast response time and high accuracy, the requirements for hardware resources increase rapidly. However, as the speaker recognition application is often implemented on mobile devices, it is necessary to maintain a low computational cost while keeping high accuracy in far-field condition. In this paper, we apply structural sparsification on time-delay neural networks (TDNN) to remove redundant structures and accelerate the execution. On our targeted hardware, our model can remove 60% of parameters and only slightly increasing equal error rate (EER) by 0.18% while our structural sparse model can achieve more than 1.5x speedup.

Shachi H Kumar, Eda Okur, Saurav Sahay et al.

With the recent advancements in AI, Intelligent Virtual Assistants (IVA) have become a ubiquitous part of every home. Going forward, we are witnessing a confluence of vision, speech and dialog system technologies that are enabling the IVAs to learn audio-visual groundings of utterances and have conversations with users about the objects, activities and events surrounding them. As a part of the 7th Dialog System Technology Challenges (DSTC7), for Audio Visual Scene-Aware Dialog (AVSD) track, We explore `topics' of the dialog as an important contextual feature into the architecture along with explorations around multimodal Attention. We also incorporate an end-to-end audio classification ConvNet, AclNet, into our models. We present detailed analysis of the experiments and show that some of our model variations outperform the baseline system presented for this task.

Mahesh Subedar, Ranganath Krishnan, Paulo Lopez Meyer et al.

Deep neural networks (DNNs) provide state-of-the-art results for a multitude of applications, but the approaches using DNNs for multimodal audiovisual applications do not consider predictive uncertainty associated with individual modalities. Bayesian deep learning methods provide principled confidence and quantify predictive uncertainty. Our contribution in this work is to propose an uncertainty aware multimodal Bayesian fusion framework for activity recognition. We demonstrate a novel approach that combines deterministic and variational layers to scale Bayesian DNNs to deeper architectures. Our experiments using in- and out-of-distribution samples selected from a subset of Moments-in-Time (MiT) dataset show a more reliable confidence measure as compared to the non-Bayesian baseline and the Monte Carlo dropout (MC dropout) approximate Bayesian inference. We also demonstrate the uncertainty estimates obtained from the proposed framework can identify out-of-distribution data on the UCF101 and MiT datasets. In the multimodal setting, the proposed framework improved precision-recall AUC by 10.2% on the subset of MiT dataset as compared to non-Bayesian baseline.

Saurav Sahay, Shachi H Kumar, Rui Xia et al.

Understanding Affect from video segments has brought researchers from the language, audio and video domains together. Most of the current multimodal research in this area deals with various techniques to fuse the modalities, and mostly treat the segments of a video independently. Motivated by the work of (Zadeh et al., 2017) and (Poria et al., 2017), we present our architecture, Relational Tensor Network, where we use the inter-modal interactions within a segment (intra-segment) and also consider the sequence of segments in a video to model the inter-segment inter-modal interactions. We also generate rich representations of text and audio modalities by leveraging richer audio and linguistic context alongwith fusing fine-grained knowledge based polarity scores from text. We present the results of our model on CMU-MOSEI dataset and show that our model outperforms many baselines and state of the art methods for sentiment classification and emotion recognition.

Tal Remez, Jonathan Huang, Matthew Brown

This paper presents a weakly-supervised approach to object instance segmentation. Starting with known or predicted object bounding boxes, we learn object masks by playing a game of cut-and-paste in an adversarial learning setup. A mask generator takes a detection box and Faster R-CNN features, and constructs a segmentation mask that is used to cut-and-paste the object into a new image location. The discriminator tries to distinguish between real objects, and those cut and pasted via the generator, giving a learning signal that leads to improved object masks. We verify our method experimentally using Cityscapes, COCO, and aerial image datasets, learning to segment objects without ever having seen a mask in training. Our method exceeds the performance of existing weakly supervised methods, without requiring hand-tuned segment proposals, and reaches 90% of supervised performance.

Saining Xie, Chen Sun, Jonathan Huang et al.

Despite the steady progress in video analysis led by the adoption of convolutional neural networks (CNNs), the relative improvement has been less drastic as that in 2D static image classification. Three main challenges exist including spatial (image) feature representation, temporal information representation, and model/computation complexity. It was recently shown by Carreira and Zisserman that 3D CNNs, inflated from 2D networks and pretrained on ImageNet, could be a promising way for spatial and temporal representation learning. However, as for model/computation complexity, 3D CNNs are much more expensive than 2D CNNs and prone to overfit. We seek a balance between speed and accuracy by building an effective and efficient video classification system through systematic exploration of critical network design choices. In particular, we show that it is possible to replace many of the 3D convolutions by low-cost 2D convolutions. Rather surprisingly, best result (in both speed and accuracy) is achieved when replacing the 3D convolutions at the bottom of the network, suggesting that temporal representation learning on high-level semantic features is more useful. Our conclusion generalizes to datasets with very different properties. When combined with several other cost-effective designs including separable spatial/temporal convolution and feature gating, our system results in an effective video classification system that that produces very competitive results on several action classification benchmarks (Kinetics, Something-something, UCF101 and HMDB), as well as two action detection (localization) benchmarks (JHMDB and UCF101-24).

Chenxi Liu, Barret Zoph, Maxim Neumann et al.

We propose a new method for learning the structure of convolutional neural networks (CNNs) that is more efficient than recent state-of-the-art methods based on reinforcement learning and evolutionary algorithms. Our approach uses a sequential model-based optimization (SMBO) strategy, in which we search for structures in order of increasing complexity, while simultaneously learning a surrogate model to guide the search through structure space. Direct comparison under the same search space shows that our method is up to 5 times more efficient than the RL method of Zoph et al. (2018) in terms of number of models evaluated, and 8 times faster in terms of total compute. The structures we discover in this way achieve state of the art classification accuracies on CIFAR-10 and ImageNet.

Ramakrishna Vedantam, Ian Fischer, Jonathan Huang et al.

It is easy for people to imagine what a man with pink hair looks like, even if they have never seen such a person before. We call the ability to create images of novel semantic concepts visually grounded imagination. In this paper, we show how we can modify variational auto-encoders to perform this task. Our method uses a novel training objective, and a novel product-of-experts inference network, which can handle partially specified (abstract) concepts in a principled and efficient way. We also propose a set of easy-to-compute evaluation metrics that capture our intuitive notions of what it means to have good visual imagination, namely correctness, coverage, and compositionality (the 3 C's). Finally, we perform a detailed comparison of our method with two existing joint image-attribute VAE methods (the JMVAE method of Suzuki et.al. and the BiVCCA method of Wang et.al.) by applying them to two datasets: the MNIST-with-attributes dataset (which we introduce here), and the CelebA dataset.

Katerina Fragkiadaki, Jonathan Huang, Alex Alemi et al.

Given a visual history, multiple future outcomes for a video scene are equally probable, in other words, the distribution of future outcomes has multiple modes. Multimodality is notoriously hard to handle by standard regressors or classifiers: the former regress to the mean and the latter discretize a continuous high dimensional output space. In this work, we present stochastic neural network architectures that handle such multimodality through stochasticity: future trajectories of objects, body joints or frames are represented as deep, non-linear transformations of random (as opposed to deterministic) variables. Such random variables are sampled from simple Gaussian distributions whose means and variances are parametrized by the output of convolutional encoders over the visual history. We introduce novel convolutional architectures for predicting future body joint trajectories that outperform fully connected alternatives \cite{DBLP:journals/corr/WalkerDGH16}. We introduce stochastic spatial transformers through optical flow warping for predicting future frames, which outperform their deterministic equivalents \cite{DBLP:journals/corr/PatrauceanHC15}. Training stochastic networks involves an intractable marginalization over stochastic variables. We compare various training schemes that handle such marginalization through a) straightforward sampling from the prior, b) conditional variational autoencoders \cite{NIPS2015_5775,DBLP:journals/corr/WalkerDGH16}, and, c) a proposed K-best-sample loss that penalizes the best prediction under a fixed "prediction budget". We show experimental results on object trajectory prediction, human body joint trajectory prediction and video prediction under varying future uncertainty, validating quantitatively and qualitatively our architectural choices and training schemes.

Michael Figurnov, Maxwell D. Collins, Yukun Zhu et al.

This paper proposes a deep learning architecture based on Residual Network that dynamically adjusts the number of executed layers for the regions of the image. This architecture is end-to-end trainable, deterministic and problem-agnostic. It is therefore applicable without any modifications to a wide range of computer vision problems such as image classification, object detection and image segmentation. We present experimental results showing that this model improves the computational efficiency of Residual Networks on the challenging ImageNet classification and COCO object detection datasets. Additionally, we evaluate the computation time maps on the visual saliency dataset cat2000 and find that they correlate surprisingly well with human eye fixation positions.

Jonathan Huang, Vivek Rathod, Chen Sun et al.

The goal of this paper is to serve as a guide for selecting a detection architecture that achieves the right speed/memory/accuracy balance for a given application and platform. To this end, we investigate various ways to trade accuracy for speed and memory usage in modern convolutional object detection systems. A number of successful systems have been proposed in recent years, but apples-to-apples comparisons are difficult due to different base feature extractors (e.g., VGG, Residual Networks), different default image resolutions, as well as different hardware and software platforms. We present a unified implementation of the Faster R-CNN [Ren et al., 2015], R-FCN [Dai et al., 2016] and SSD [Liu et al., 2015] systems, which we view as "meta-architectures" and trace out the speed/accuracy trade-off curve created by using alternative feature extractors and varying other critical parameters such as image size within each of these meta-architectures. On one extreme end of this spectrum where speed and memory are critical, we present a detector that achieves real time speeds and can be deployed on a mobile device. On the opposite end in which accuracy is critical, we present a detector that achieves state-of-the-art performance measured on the COCO detection task.

Jonathan Huang, Kevin Murphy

We present a generative model of images based on layering, in which image layers are individually generated, then composited from front to back. We are thus able to factor the appearance of an image into the appearance of individual objects within the image --- and additionally for each individual object, we can factor content from pose. Unlike prior work on layered models, we learn a shape prior for each object/layer, allowing the model to tease out which object is in front by looking for a consistent shape, without needing access to motion cues or any labeled data. We show that ordinary stochastic gradient variational bayes (SGVB), which optimizes our fully differentiable lower-bound on the log-likelihood, is sufficient to learn an interpretable representation of images. Finally we present experiments demonstrating the effectiveness of the model for inferring foreground and background objects in images.

Vignesh Ramanathan, Jonathan Huang, Sami Abu-El-Haija et al.

Multi-person event recognition is a challenging task, often with many people active in the scene but only a small subset contributing to an actual event. In this paper, we propose a model which learns to detect events in such videos while automatically "attending" to the people responsible for the event. Our model does not use explicit annotations regarding who or where those people are during training and testing. In particular, we track people in videos and use a recurrent neural network (RNN) to represent the track features. We learn time-varying attention weights to combine these features at each time-instant. The attended features are then processed using another RNN for event detection/classification. Since most video datasets with multiple people are restricted to a small number of videos, we also collected a new basketball dataset comprising 257 basketball games with 14K event annotations corresponding to 11 event classes. Our model outperforms state-of-the-art methods for both event classification and detection on this new dataset. Additionally, we show that the attention mechanism is able to consistently localize the relevant players.

Chris Piech, Jonathan Spencer, Jonathan Huang et al.

Knowledge tracing---where a machine models the knowledge of a student as they interact with coursework---is a well established problem in computer supported education. Though effectively modeling student knowledge would have high educational impact, the task has many inherent challenges. In this paper we explore the utility of using Recurrent Neural Networks (RNNs) to model student learning. The RNN family of models have important advantages over previous methods in that they do not require the explicit encoding of human domain knowledge, and can capture more complex representations of student knowledge. Using neural networks results in substantial improvements in prediction performance on a range of knowledge tracing datasets. Moreover the learned model can be used for intelligent curriculum design and allows straightforward interpretation and discovery of structure in student tasks. These results suggest a promising new line of research for knowledge tracing and an exemplary application task for RNNs.

Chris Piech, Jonathan Huang, Andy Nguyen et al.

Providing feedback, both assessing final work and giving hints to stuck students, is difficult for open-ended assignments in massive online classes which can range from thousands to millions of students. We introduce a neural network method to encode programs as a linear mapping from an embedded precondition space to an embedded postcondition space and propose an algorithm for feedback at scale using these linear maps as features. We apply our algorithm to assessments from the Code.org Hour of Code and Stanford University's CS1 course, where we propagate human comments on student assignments to orders of magnitude more submissions.

Jonathan Malmaud, Jonathan Huang, Vivek Rathod et al.

We present a novel method for aligning a sequence of instructions to a video of someone carrying out a task. In particular, we focus on the cooking domain, where the instructions correspond to the recipe. Our technique relies on an HMM to align the recipe steps to the (automatically generated) speech transcript. We then refine this alignment using a state-of-the-art visual food detector, based on a deep convolutional neural network. We show that our technique outperforms simpler techniques based on keyword spotting. It also enables interesting applications, such as automatically illustrating recipes with keyframes, and searching within a video for events of interest.

Jonathan Huang, Ashish Kapoor, Carlos Guestrin

Distributions over rankings are used to model data in a multitude of real world settings such as preference analysis and political elections. Modeling such distributions presents several computational challenges, however, due to the factorial size of the set of rankings over an item set. Some of these challenges are quite familiar to the artificial intelligence community, such as how to compactly represent a distribution over a combinatorially large space, and how to efficiently perform probabilistic inference with these representations. With respect to ranking, however, there is the additional challenge of what we refer to as human task complexity users are rarely willing to provide a full ranking over a long list of candidates, instead often preferring to provide partial ranking information. Simultaneously addressing all of these challenges i.e., designing a compactly representable model which is amenable to efficient inference and can be learned using partial ranking data is a difficult task, but is necessary if we would like to scale to problems with nontrivial size. In this paper, we show that the recently proposed riffled independence assumptions cleanly and efficiently address each of the above challenges. In particular, we establish a tight mathematical connection between the concepts of riffled independence and of partial rankings. This correspondence not only allows us to then develop efficient and exact algorithms for performing inference tasks using riffled independence based represen- tations with partial rankings, but somewhat surprisingly, also shows that efficient inference is not possible for riffle independent models (in a certain sense) with observations which do not take the form of partial rankings. Finally, using our inference algorithm, we introduce the first method for learning riffled independence based models from partially ranked data.

Chris Piech, Jonathan Huang, Zhenghao Chen et al.

In massive open online courses (MOOCs), peer grading serves as a critical tool for scaling the grading of complex, open-ended assignments to courses with tens or hundreds of thousands of students. But despite promising initial trials, it does not always deliver accurate results compared to human experts. In this paper, we develop algorithms for estimating and correcting for grader biases and reliabilities, showing significant improvement in peer grading accuracy on real data with 63,199 peer grades from Coursera's HCI course offerings --- the largest peer grading networks analysed to date. We relate grader biases and reliabilities to other student factors such as student engagement, performance as well as commenting style. We also show that our model can lead to more intelligent assignment of graders to gradees.

Jonathan Huang, Ashish Kapoor, Carlos E. Guestrin

Distributions over rankings are used to model data in various settings such as preference analysis and political elections. The factorial size of the space of rankings, however, typically forces one to make structural assumptions, such as smoothness, sparsity, or probabilistic independence about these underlying distributions. We approach the modeling problem from the computational principle that one should make structural assumptions which allow for efficient calculation of typical probabilistic queries. For ranking models, "typical" queries predominantly take the form of partial ranking queries (e.g., given a user's top-k favorite movies, what are his preferences over remaining movies?). In this paper, we argue that riffled independence factorizations proposed in recent literature [7, 8] are a natural structural assumption for ranking distributions, allowing for particularly efficient processing of partial ranking queries.