Search for: All records

Multi-Channel Imaging (MCI) contains an array of challenges for encoding useful feature representations not present in traditional images. For example, images from two different satellites may both contain RGB channels, but the remaining channels can be different for each imaging source. Thus, MCI models must support a variety of channel configurations at test time. Recent work has extended traditional visual encoders for MCI, such as Vision Transformers (ViT), by supplementing pixel information with an encoding representing the channel configuration. However, these methods treat each channel equally, i.e., they do not consider the unique properties of each channel type, which can result in needless and potentially harmful redundancies in the learned features. For example, if RGB channels are always present, the other channels can focus on extracting information that cannot be captured by the RGB channels. To this end, we propose DiChaViT, which aims to enhance the diversity in the learned features of MCI-ViT models. This is achieved through a novel channel sampling strategy that encourages the selection of more distinct channel sets for training. Additionally, we employ regularization and initialization techniques to increase the likelihood that new information is learned from each channel. Many of our improvements are architecture agnostic and can be incorporated into new architectures as they are developed. Experiments on both satellite and cell microscopy datasets, CHAMMI, JUMP-CP, and So2Sat, report DiChaViT yields a 1.5 - 5.0% gain over the state-of-the-art. Our code is publicly available at https://github.com/chaudatascience/diversechannelvit. 

Most neural networks assume that input images have a fixed number of channels (three for RGB images). However, there are many settings where the number of channels may vary, such as microscopy images where the number of channels changes depending on instruments and experimental goals. Yet, there has not been a systemic attempt to create and evaluate neural networks that are invariant to the number and type of channels. As a result, trained models remain specific to individual studies and are hardly reusable for other microscopy settings. In this paper, we present a benchmark for investigating channel-adaptive models in microscopy imaging, which consists of 1) a dataset of varied-channel single-cell images, and 2) a biologically relevant evaluation framework. In addition, we adapted several existing techniques to create channel-adaptive models and compared their performance on this benchmark to fixed-channel, baseline models. We find that channel-adaptive models can generalize better to out-of-domain tasks and can be computationally efficient. We contribute a curated dataset and an evaluation API to facilitate objective comparisons in future research and applications. 

Understanding emotions that people express during large-scale crises helps inform policy makers and first responders about the emotional states of the population as well as provide emotional support to those who need such support. We present CovidEmo, a dataset of ~3,000 English tweets labeled with emotions and temporally distributed across 18 months. Our analyses reveal the emotional toll caused by COVID-19, and changes of the social narrative and associated emotions over time. Motivated by the time-sensitive nature of crises and the cost of large-scale annotation efforts, we examine how well large pre-trained language models generalize across domains and timeline in the task of perceived emotion prediction in the context of COVID-19. Our analyses suggest that cross-domain information transfers occur, yet there are still significant gaps. We propose semi-supervised learning as a way to bridge this gap, obtaining significantly better performance using unlabeled data from the target domain. 

A paper by Zhang et al. in 2001, “On the Constancy of Internet Path Properties” [1] examined the constancy of end- to-end packet loss, latency, and throughput using a modest set of hosts deployed in the Internet. In the time since that work, the Internet has changed dramatically, including the flattening of the autonomous system hierarchy and increased deployment of IPv6, among other developments. In this paper, we investigate the constancy of end-to-end Internet latency, revisiting findings of the earlier study. We use latency measurements from RIPE Atlas, choosing a set of 124 anchors with broad geographic distribution and drawn from 112 distinct autonomous systems. The earlier work of Zhang et al. relies on changepoint detection methods to identify mathematically constant time periods. We reimplement the two methods described in that earlier work and use them on the RIPE Atlas latency measurements. We also use a recently- published method (HMM-HDP) that has direct support in a RIPE Atlas API. Comparing the three changepoint detection methods, we find that the two methods used in the earlier work may miss many changepoints caused by common level-shift events. Overall, we find that the recently proposed HMM-HDP method performs substantially better. Moreover, we find that delay spikes—as defined by the earlier work—are an order of magnitude less prevalent than 20 years ago. We also find that maximum change- free regions (CFRs) along paths that we observe in today’s Internet are substantially longer than what was observed in 2001, regardless of the changepoint detection method used. In particular, the 50th percentile maximum CFR was on the order of 30 minutes in the earlier study, but our analysis reveals it to be on the order of 3 days or longer. Moreover, we find that CFR durations appear to have steadily increased over the past 5 years.