Search Available Technologies
Search by category
Displaying Available Technologies results for Computing
CORE FACILITY SCHEDULING AND BILLING SYSTEM
Core facilities rely on their internal operations to function smoothly. Yet, in a survey of 156 core facilities, over 50 hours per month were spent on managing business that included tracking requests, staff, inventory, and billing.
The University of Utah’s core facilities administrators have developed a billing and scheduling laboratory inventory management system (LIMS). The LIMS system enables control, scheduling, and billing services across multiple departments. This system facilitates coordinated use of instruments and resources. Email notifications are sent to appropriate users following major system events, and all records are stored and easily accessible to administrators.
HEART RATE REGULATION APP
The majority of heart rate regulation apps require manual adjustment of desired heart rates, leading to over- and under-exertion.
University of Utah researchers have created an application and wearable for automatic, personalized, high accuracy heart rate regulation. The user inputs their desired heart rate plan, and a wearable heart rate sensor provides real-time bio-feedback to influence audible cues produced by the application. This adaptive audible signal gives the runner a live, calculated cadence by which to adjust their pace.
MISTREAM: LIVE CROWD STREAMING
Live-streaming is growing in popularity with the rise of applications like Facebook Live, YouTube, Meerkat and Periscope. Yet, live-streaming has limitations: video streaming often requires large amounts of bandwidth, resulting in lower quality video/audio and loss of signal.
mIStream splits streaming cellular data to numerous forwarder cells which then transmit the data via multiple cellular paths. A gatherer then combines the multiple streams of data to seamlessly recompile the original stream. This creates a more robust signal network, allowing high-quality, higher-bandwidth streaming despite individual cellular data limitations and fluctuations. Having forwarder nodes on multiple cellular networks reduces dead zones and improves the overall throughput.
POWER GRID CONTINUOUS-TIME SCHEDULING, PRICING, & STORAGE OPTIMIZATION
The lag between demand spikes and energy production, a common situation referred to as a ramping scarcity event, results in higher consumer prices, overburdened energy grids, and higher costs to utility companies.
U of U researchers have developed a software suite composed of multiple algorithms that generate continuous-time estimates of the most efficacious strategies for energy pricing, generation, scheduling, and storage. Continuous-time demand optimization is based on marginal pricing, flexible loads, and power generation ramping trajectories. These optimization algorithms have been created to avoid ramping scarcity events, more accurately predict consumer prices, and off-load overburdened energy grids to available energy storage devices.
LARGE DATABASE QUERY OPTIMIZATION
Large local database queries are critical and often time-consuming. In relational databases, for example, joins are the most costly operations, but provide the most valuable information.
A software suite has been developed that speeds database queries by an order-of-magnitude. The suite includes software designed for local database crawling that enables users to deploy a Google-esque query engine over their data almost instantaneously. In addition, STORM (spatio-temporal online reasoning and management) software allows for online analytics of multi-dimensional data, incorporating machine learning into database query analysis. The final software piece speeds join queries in relational databases, facilitating interactive data analytics at user-friendly speeds.
DEBUGGING MACHINE LEARNING SYSTEMS
Developers are unable to debug, optimize, or even understand the processes that generate machine learning outputs due to the complexity of machine learning systems. This “black box” problem of machine learning poses significant limitations to widespread implementation of artificial intelligence.
A University of Utah researcher has developed software for “cracking open” the machine learning black box. This software integrates systems analysis with machine learning to force machine learning systems to express a linear dynamic systems equation. The linear equation is tested in conjunction with a machine learning technique that tests nonlinear relationships. The combination of these techniques enables mapping of machine learning internal processes over time.
ENVIRONMENTAL DISPERSION SOFTWARE
Continuous commercial and residential developments have negatively impacted the environment by constraining resources, while increasing noise and air pollution. Computational dynamics solvers simulate the interaction between infrastructure and the environment, allowing civil engineers to understand environmental variables that impact design. Use of these simulations, however, is limited by time constraints because model development often takes multiple days.
The Environmental Dispersion Software evaluates various design scenarios by rapidly simulating relevant climate variables. The software uses traditional central processing and new graphics processing units to produce solutions based on buildings, vegetation, and other parts of a city more rapidly than traditional computational dynamics solvers. The Environmental Dispersion Software computes wind speed/direction, turbulence, air temperature, humidity, and atmospheric radiation in a given area. The software also evaluates how these factors, as well as dispersion and deposition of particle contaminants, will affect buildings and vegetation. The software can be used to evaluate urban designs for optimal air quality, energy use, and environmental impact.
SPELL & DEEPLOG
System logs record system states at critical points to help debug failures and promote system stability. Analyzing system logs to detect irregularities establishes more secure and trustworthy systems. Typical log parsing software provides offline, batch processing of raw files, but many applications require constant monitoring not provided by offline methods.
Spell, an online streaming method, parses system event logs to dynamically extract log patterns and maintain a set of discovered message types. DeepLog utilizes Long Short-Term Memory (LSTM) to model a system log as a natural language sequence that automatically learns log patterns. DeepLog detects anomalies when log patterns deviate from the model trained from log data under normal execution. When an anomaly is detected, users can diagnose and perform root cause analysis immediately, thereby increasing system security.
SECRET KEY EXCHANGE FOR WIRELESS COMMUNICATIONS
Secret key establishment between two entities is a fundamental requirement for private communication. The most common method for establishing a secret key is by using public key cryptography. The public key method, however, is vulnerable to security breaches, consumes significant amount of computing resources, has been relatively slow, and requires a third party authentication service.
The proposed method represents a fundamental advancement in secret key exchange for wireless communications by avoiding use of a public key. It allows for the exchange of a random secret key between two transceivers by measuring particular bio-directional properties of the channel, which cannot be read at a third location. The properties may include measurements while moving a transceiver to different locations to observe the secret key. The method takes advantage of the changing space-time wireless channel in order to generate a rich and robust key which can be shared on a wireless link without communicating the secret key.