arXiv subjects

Jonathan Huang

Publications and source records attributed to Jonathan Huang.

At least 19 recordsLinked to original sources

arXiv2025

From Scope to Script: An Automated Report Generation Model for Gastrointestinal Endoscopy

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.

cs.CV↗

arXiv2025

MALT Diffusion: Memory-Augmented Latent Transformers for Any-Length Video Generation

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.

cs.CV↗

arXiv2024

Learning Complex Non-Rigid Image Edits from Multimodal Conditioning

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.

cs.CV↗

arXiv2024

Principles of Visual Tokens for Efficient Video Understanding

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.

cs.CV↗

arXiv2024

Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think

Recent studies have shown that the denoising process in (generative) diffusion models can induce meaningful (discriminative) representations inside the model, though the quality of these representations still lags behind those learned through recent self-supervised learning methods. We argue that one main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations. Moreover, training can be made easier by incorporating high-quality external visual representations, rather than relying solely on the diffusion models to learn them independently. We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders. The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs. For instance, our method can speed up SiT training by over 17.5$\times$, matching the performance (without classifier-free guidance) of a SiT-XL model trained for 7M steps in less than 400K steps. In terms of final generation quality, our approach achieves state-of-the-art results of FID=1.42 using classifier-free guidance with the guidance interval.

cs.CV↗

arXiv2024

Tree-D Fusion: Simulation-Ready Tree Dataset from Single Images with Diffusion Priors

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.

cs.CV↗

arXiv2024

Visually Consistent Hierarchical Image Classification

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.

cs.CV↗

arXiv2023

VideoPoet: A Large Language Model for Zero-Shot Video Generation

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/

cs.CV↗

arXiv2023

Text and Click inputs for unambiguous open vocabulary instance segmentation

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.

cs.CV↗

arXiv2023

Optimizing ViViT Training: Time and Memory Reduction for Action Recognition

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.

cs.CV↗

arXiv2023

DaTaSeg: Taming a Universal Multi-Dataset Multi-Task Segmentation Model

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.

cs.CV↗

arXiv2023

Learning to Detect Novel and Fine-Grained Acoustic Sequences Using Pretrained Audio Representations

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.

cs.SD↗

arXiv2022

Partisan US News Media Representations of Syrian Refugees

We investigate how representations of Syrian refugees (2011-2021) differ across US partisan news outlets. We analyze 47,388 articles from the online US media about Syrian refugees to detail differences in reporting between left- and right-leaning media. We use various NLP techniques to understand these differences. Our polarization and question answering results indicated that left-leaning media tended to represent refugees as child victims, welcome in the US, and right-leaning media cast refugees as Islamic terrorists. We noted similar results with our sentiment and offensive speech scores over time, which detail possibly unfavorable representations of refugees in right-leaning media. A strength of our work is how the different techniques we have applied validate each other. Based on our results, we provide several recommendations. Stakeholders may utilize our findings to intervene around refugee representations, and design communications campaigns that improve the way society sees refugees and possibly aid refugee outcomes.

cs.SI↗

arXiv2021

REFINE2: A tool to evaluate real-world performance of machine-learning based effect estimators for molecular and clinical studies

Data-adaptive (machine learning-based) effect estimators are increasingly popular to reduce bias in high-dimensional bioinformatic and clinical studies (e.g. real-world data, target trials, -omic discovery). Their relative statistical efficiency (high power) is particularly invaluable in these contexts since sample sizes are often limited due to practical and cost concerns. However, these methods are subject to technical limitations that are dataset specific and involve computational trade-offs. Thus, it is challenging for analysts to identify when such methods may offer benefits or select amongst statistical methods. We present extensive simulation studies of several cutting-edge estimators, evaluating both performance and computation time. Critically, rather than use arbitrary simulation data, we generate synthetic datasets mimicking the observed data structure (plasmode simulation) of a real molecular epidemiologic cohort. We find that machine learning approaches may not always be indicated in such data settings, but that performance is highly context dependent. We present a user-friendly Shiny app REFINE2 (Realistic Evaluations of Finite sample INference using Efficient Estimators) that enables analysts to simulate synthetic data from their own datasets and directly evaluate the performance of several cutting-edge algorithms in those settings. This tool may greatly facilitate the proper selection and implementation of machine-learning-based effect estimators in bioinformatic and clinical study contexts.

stat.ME↗

arXiv2021

Local Metrics for Multi-Object Tracking

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.

cs.CV↗

arXiv2021

The surprising impact of mask-head architecture on novel class segmentation

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.

cs.CV↗

arXiv2020

The implications of outcome truncation in reproductive medicine RCTs: a simulation platform for trialists and simulation study

Randomised controlled trials in reproductive medicine are often subject to outcome truncation, where study outcomes are only defined in a subset of participants. Examples include birthweight (measurable only in the subgroup of participants who give birth) and miscarriage (which can only occur in participants who become pregnant). These are typically analysed by making a comparison between treatment arms within the subgroup (comparing birthweights in the subgroup who gave birth, or miscarriages in the subgroup who became pregnant). However, this approach does not represent a randomised comparison when treatment influences the probability of being observed (i.e. survival). The practical implications of this for reproductive trials are unclear. We developed a simulation platform to investigate the implications of outcome truncation for reproductive medicine trials. We used this to perform a simulation study, in which we considered the bias, Type 1 error, coverage, and precision of standard statistical analyses for truncated continuous and binary outcomes. Increasing treatment effect on the intermediate variable, strength of confounding between the intermediate and outcome variables, and interactions between treatment and confounder were found to adversely affect performance. However, within parameter ranges we would consider to be more realistic, the adverse effects were generally not drastic. For binary outcomes, the study highlighted that outcome truncation may lead to none of the participants in a study arm experiencing the outcome event. This was found to have severe consequences for inferences, and this may have implications for meta-analysis.

stat.AP↗

arXiv2020

PERF-Net: Pose Empowered RGB-Flow Net

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.

cs.CV↗