LLM Comparator: Interactive Analysis of Side-by-Side Evaluation of Large Language Models.

Evaluating large language models (LLMs) presents unique challenges. While automatic side-by-side evaluation, also known as LLM-as-a-judge, has become a promising solution, model developers and researchers face difficulties with scalability and interpretability when analyzing these evaluation outcomes. To address these challenges, we introduce LLM Comparator, a new visual analytics tool designed for side-by-side evaluations of LLMs.

Calcitonin receptor signaling in nucleus accumbens D1R- and D2R-expressing medium spiny neurons bidirectionally alters opioid taking in male rats.

The high rates of relapse associated with current medications used to treat opioid use disorder (OUD) necessitate research that expands our understanding of the neural mechanisms regulating opioid taking to identify molecular substrates that could be targeted by novel pharmacotherapies to treat OUD. Recent studies show that activation of calcitonin receptors (CTRs) is sufficient to reduce the rewarding effects of addictive drugs in rodents. However, the role of central CTR signaling in opioid-mediated behaviors has not been studied.

DendroMap: Visual Exploration of Large-Scale Image Datasets for Machine Learning with Treemaps.

In this paper, we present DendroMap, a novel approach to interactively exploring large-scale image datasets for machine learning (ML). ML practitioners often explore image datasets by generating a grid of images or projecting high-dimensional representations of images into 2-D using dimensionality reduction techniques (e.g.

A novel dual agonist of glucagon-like peptide-1 receptors and neuropeptide Y2 receptors attenuates fentanyl taking and seeking in male rats.

There has been a dramatic increase in illicit fentanyl use in the United States over the last decade. In 2018, more than 31,000 overdose deaths involved fentanyl or fentanyl analogs, highlighting an urgent need to identify effective treatments for fentanyl use disorder. An emerging literature shows that glucagon-like peptide-1 receptor (GLP-1R) agonists attenuate the reinforcing efficacy of drugs of abuse.

CNN Explainer: Learning Convolutional Neural Networks with Interactive Visualization.

Deep learning's great success motivates many practitioners and students to learn about this exciting technology. However, it is often challenging for beginners to take their first step due to the complexity of understanding and applying deep learning. We present CNN Explainer, an interactive visualization tool designed for non-experts to learn and examine convolutional neural networks (CNNs), a foundational deep learning model architecture.

Activation of GLP-1 receptors attenuates oxycodone taking and seeking without compromising the antinociceptive effects of oxycodone in rats.

Despite the effectiveness of current medications to treat opioid use disorder, there is still a high rate of relapse following detoxification. Thus, there is critical need for innovative studies aimed at identifying novel neurobiological mechanisms that could be targeted to treat opioid use disorder. A growing body of preclinical evidence indicates that glucagon-like peptide-1 (GLP-1) receptor agonists reduce drug reinforcement.

GAN Lab: Understanding Complex Deep Generative Models using Interactive Visual Experimentation.

Recent success in deep learning has generated immense interest among practitioners and students, inspiring many to learn about this new technology. While visual and interactive approaches have been successfully developed to help people more easily learn deep learning, most existing tools focus on simpler models. In this work, we present GAN Lab, the first interactive visualization tool designed for non-experts to learn and experiment with Generative Adversarial Networks (GANs), a popular class of complex deep learning models.

Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers.

Deep learning has recently seen rapid development and received significant attention due to its state-of-the-art performance on previously-thought hard problems. However, because of the internal complexity and nonlinear structure of deep neural networks, the underlying decision making processes for why these models are achieving such performance are challenging and sometimes mystifying to interpret. As deep learning spreads across domains, it is of paramount importance that we equip users of deep learning with tools for understanding when a model works correctly, when it fails, and ultimately how to improve its performance.

Chronodes: Interactive Multifocus Exploration of Event Sequences.

The advent of mobile health (mHealth) technologies challenges the capabilities of current visualizations, interactive tools, and algorithms. We present Chronodes, an interactive system that unifies data mining and human-centric visualization techniques to support explorative analysis of longitudinal mHealth data. Chronodes extracts and visualizes frequent event sequences that reveal chronological patterns across multiple participant timelines of mHealth data.

mHealth Visual Discovery Dashboard.

We present Discovery Dashboard, a visual analytics system for exploring large volumes of time series data from mobile medical field studies. Discovery Dashboard offers interactive exploration tools and a data mining motif discovery algorithm to help researchers formulate hypotheses, discover trends and patterns, and ultimately gain a deeper understanding of their data. Discovery Dashboard emphasizes user freedom and flexibility during the data exploration process and enables researchers to do things previously challenging or impossible to do - in the web-browser and in real time.

ACTIVIS: Visual Exploration of Industry-Scale Deep Neural Network Models.

While deep learning models have achieved state-of-the-art accuracies for many prediction tasks, understanding these models remains a challenge. Despite the recent interest in developing visual tools to help users interpret deep learning models, the complexity and wide variety of models deployed in industry, and the large-scale datasets that they used, pose unique design challenges that are inadequately addressed by existing work. Through participatory design sessions with over 15 researchers and engineers at Facebook, we have developed, deployed, and iteratively improved ACTIVIS, an interactive visualization system for interpreting large-scale deep learning models and results.

GLO-STIX: Graph-Level Operations for Specifying Techniques and Interactive eXploration.

The field of graph visualization has produced a wealth of visualization techniques for accomplishing a variety of analysis tasks. Therefore analysts often rely on a suite of different techniques, and visual graph analysis application builders strive to provide this breadth of techniques. To provide a holistic model for specifying network visualization techniques (as opposed to considering each technique in isolation) we present the Graph-Level Operations (GLO) model.

GLO-STIX: Graph-Level Operations for Specifying Techniques and Interactive eXploration.

The field of graph visualization has produced a wealth of visualization techniques for accomplishing a variety of analysis tasks. Therefore analysts often rely on a suite of different techniques, and visual graph analysis application builders strive to provide this breadth of techniques. To provide a holistic model for specifying network visualization techniques (as opposed to considering each technique in isolation) we present the () model.

MMap: Fast Billion-Scale Graph Computation on a PC via Memory Mapping.

Graph computation approaches such as GraphChi and TurboGraph recently demonstrated that a single PC can perform efficient computation on billion-node graphs. To achieve high speed and scalability, they often need sophisticated data structures and memory management strategies. We propose a minimalist approach that forgoes such requirements, by leveraging the fundamental (MMap) capability found on operating systems.

Towards Scalable Graph Computation on Mobile Devices.

Mobile devices have become increasingly central to our everyday activities, due to their portability, multi-touch capabilities, and ever-improving computational power. Such attractive features have spurred research interest in leveraging mobile devices for computation. We explore a novel approach that aims to use a mobile device to perform scalable graph computation on large graphs that do not fit in the device's limited main memory, opening up the possibility of performing on-device analysis of large datasets, without relying on the cloud.

Understanding variations in pediatric asthma care processes in the emergency department using visual analytics.

Health care delivery processes consist of complex activity sequences spanning organizational, spatial, and temporal boundaries. Care is human-directed so these processes can have wide variations in cost, quality, and outcome making systemic care process analysis, conformance testing, and improvement challenging. We designed and developed an interactive visual analytic process exploration and discovery tool and used it to explore clinical data from 5784 pediatric asthma emergency department patients.

High-affinity renal lead-binding proteins in environmentally-exposed humans.

Chronic low level lead (Pb) exposure is associated with decrements in renal function in humans, but the molecular mechanisms underlying toxicity are not understood. We investigated cytosolic Pb-binding proteins (PbBP) in kidney of environmentally-exposed humans to identify molecular targets of Pb and elucidate mechanisms of toxicity. This study is unique in that it localized PbBPs based on physiologic Pb that was bound in vivo.

Lead-binding proteins in brain tissue of environmentally lead-exposed humans.

This study reports the partial purification and characterization of cytosolic lead binding proteins (PbBPs) in human brain tissue of environmentally Pb-exposed subjects. The isolated proteins were initially characterized based upon the presence of endogenously associated Pb. Following partial purification (Sephadex G-75 and A-25 DEAE anion-exchange chromatography), the isolated PbBPs (contained within a single DEAE peak) showed a single class of high affinity binding sites with an apparent Kd of 10(-9) M, based upon competition assays using radioactive 203Pb and Hill and Scatchard analysis.

Role of lead-binding proteins in renal cancer.

High-dose lead exposure in rodents has been shown to produce pathognomonic lead intranuclear inclusion bodies and to result in an increased incidence of renal adenocarcinomas. Studies from this laboratory and others have demonstrated the presence of high-affinity renal lead-binding proteins in rat kidneys which act as tissue sinks for lead at low dose levels. Cell-free nuclear translocation studies have shown that these molecules are capable of facilitating the intranuclear movement of lead and that they are associated with chromatin.

Detection of DNA point mutations with DNA mismatch repair enzymes.

We have developed a simple and reliable procedure to screen gene mutations using DNA mismatch repair (MR) specific mut Y enzyme of Escherichia coli and thymidine DNA glycosylase from HeLa cells. The mut Y enzyme cleaves A of G/A mismatches in DNA duplex and thymidine glycosylase cleaves T at G/T mismatches. Previously, we showed the determination of G:C-->T:A mutations in the N-ras gene in two human tumor samples with mut Y G/A MR enzyme.

Implications of lead binding proteins for risk assessment of lead exposure.

Lead-binding proteins have previously been isolated from rat and human target tissues. These molecules have shown to possess molecular masses in the general range of 10,000-30,000 daltons. The proteins are acidic in nature and rich in aspartic and glutamic amino acid residues.

Methods for improving tissue culture of human tracheo-bronchial epithelium obtained at autopsy.

Human tracheo-bronchial epithelium obtained from autopsy, surgery, and organ donation will have areas of both viable and non-viable cells. It is important in the initial establishment of epithelial explant and cell cultures that injured, non-viable mucosal epithelium not be used for the cultures. Autopsy cases selected for culture should initially be chosen on the basis of a shorter post mortem interval and cause of death in order to increase the rate of successful culture.

Characterization of an in vitro system of human renal papillary collecting duct cells.

We have developed an in vitro model of human papillary collecting duct cells isolated from cadaver kidneys using methods similar to those we previously reported for the isolation of human proximal tubule cells. To date we have isolated papillary collecting duct cells from 100 normal human kidneys. Papillae were dissected and digested in Cellgro containing 400 U/ml collagenase.