Applications

Latest Innovations
DNA-Based Data Storage Using Portable Sequencer Dr. Milenkovic has developed a new method for error-free random access data storage through DNA that uses a portable nanopore sequencer. The error correction methods used... Read More Dr. Milenkovic has developed a new method for error-free random access data storage through DNA that uses a portable nanopore sequencer. The error correction methods used in this scheme result in error-free reconstructions of the synthetic DNA strand. This system has an information rate of 0.83 and a storage density of 1.07 x 10^23 bytes/gram.
Barcode and Software for Identifying Insects in Digital Images Researchers working with the lab of Gene Robinson have developed a custom bar code called "bcode" with dimension of 2.1mmX2.1mm which can easily be put on insects and then... Read More Researchers working with the lab of Gene Robinson have developed a custom bar code called "bcode" with dimension of 2.1mmX2.1mm which can easily be put on insects and then tracked through a custom computer vision software. The bcodes can be printed cheaply using office printer at 600dpi.
Native DNA Based Storage and Computing via Single Strand Nicking Prof. Milenkovic from the University of Illinois at Urbana-Champaign has developed a new method for DNA based storage of information. In... Read More Prof. Milenkovic from the University of Illinois at Urbana-Champaign has developed a new method for DNA based storage of information. In this method information is first converted into positive integers based on which single strand of the DNA is nicked at defined positions. This information can be accessed through immunoprecipitation and sequencing of the DNA. This method allows for random access as well as computation of data.
Deep Learning for Fashion Compatibility and Style Drs Ranjitha Kumar, David Forsyth, and their graduate students have developed a interactive chatbot that provides fashion recommendations based on user input and clothing... Read More Drs Ranjitha Kumar, David Forsyth, and their graduate students have developed a interactive chatbot that provides fashion recommendations based on user input and clothing. The chatbot uses a neural regressor trained on thousands of examples of fashion and text to provide outfit recommendations. Users can specify the context such as outdoor picnics, formal wear, business, etc. and receive recommendations based on their existing wardrobe.
QuickStop –The Quickest Misinformation Detection Algorithm Dr. Xiaohan Kang from the University of IL, in collaboration with inventors from Arizona State University and Carnege Mellon University, has developed QuickStop, an... Read More Dr. Xiaohan Kang from the University of IL, in collaboration with inventors from Arizona State University and Carnege Mellon University, has developed QuickStop, an algorithm capable of quick misinformation detection. Application This technology has applications in stopping the spread of misinformation,or “fake news”, on social platforms or news media before it can cause harm to the public. Benefits As the name would suggest, QuickStop is quicker than other misinformation technologies while maintaining higher accuracy than its competitors. QuickStop can detect fake news after 15 observations while other technologies need up to 500 observations. It is capable of faster speed and greater accuracy because of its dynamic nature. QuickStop calculates the minimum number of observations to determine whether information is real news or fake news.
SemiSynBio: An On-Chip Nanoscale Storage System Using Chimeric DNA Dr. Milenkovic from the University of Illinois has developed an on-chip integrated nanosytem for writing, storage, access, and reading of large data volumes that utilizes... Read More Dr. Milenkovic from the University of Illinois has developed an on-chip integrated nanosytem for writing, storage, access, and reading of large data volumes that utilizes inexpensive chimeric DNA, scalable grids of semiconductor nanotubes, and specialized nanopore sequencers. Using an M13 template for ease-of-synthesis and modified DNA bases to expand the potential alphabet, this system of synthesis for DNA-based data storage is inexpensive and flexible. In addition, selectively addressable semiconductor nanotubes are incorporated into an on-chip nanopore sequencer enabling selective storage/querying/retrieval of information in DNA-based data storage. This development represents a significant advance towards end-to-end automation of DNA-based data storage. Primary Application: Data Storage and Archiving Benefit: Provides a mechanism for selective storage and reading of DNA library, basis for data storage architecture
ωLog: Transparent Interpretation and Integration of Layered Software Architecture Event Streams Prof. Bates at the University of Illinois has developed ωLog, a software application which collects application context through analysis of event logs, and integrates that... Read More Prof. Bates at the University of Illinois has developed ωLog, a software application which collects application context through analysis of event logs, and integrates that information into whole-system provenance. Through binary analysis, ωLog determines application logging behavior, associates events at the application level with events at the system level, and based on user queries, generates a concise, semantically-rich, execution-partitioned provenance graph. ωLog addresses limitations in overhead costs, with an average of 12% runtime overhead, and does not require instrumentation. Additionally, ωLog provides forensically-relevant semantic information that other provenance systems cannot provide. ωLog provides system administrators an unprecedented wealth of information for tracing suspicious activity to the root cause, which is particularly useful in identifying and responding to Advanced Persistent Threats. Applications Security Information and Event Management (SIEM), cybersecurity
Image Processing and Error-Correction in DNA via Inpainting and Filtering Dr. Olgica Milenkovic and Dr. Charles Schroeder at the University of Illinois has developed an error-correction method for DNA-based data storage of massive image datasets. Previous methods for... Read More Dr. Olgica Milenkovic and Dr. Charles Schroeder at the University of Illinois has developed an error-correction method for DNA-based data storage of massive image datasets. Previous methods for addressing errors use Reed-Solomon codes at both the individual oligo and pool of oligo level to reconstruct missing strings from redundantly encoded oligos. This method uses deep learning techniques to utilize natural redundancy present in the images to perform inpainting on the missing pixels, reconstructing the missing information, and avoiding the requirement for synthesizing and sequencing redundancy in the oligos.
Food Flows between Counties in the United States Dr. Megan Konar has developed an algorithm for mapping food flows between counties in the United States based on FAF data. This algorithm downscales FAF data to the... Read More Dr. Megan Konar has developed an algorithm for mapping food flows between counties in the United States based on FAF data. This algorithm downscales FAF data to the county level to provide a spatially-detailed map of food flows. By building on previous models with known network properties of food flows and the principle of mass balance, this algorithm improves spatial detail and accuracy of food flow mapping. Mapping flows of commodity with higher resolution enables detailed lifecycle and footprint assessments.This technology uses US Census Bureau data to provide realistic and high-resolution estimations of food flows between counties, providing information about distribution of food through complex supply chains, which underpins food security and has implications for the environment.
Forecasting 3D Movement While machines that analyze movement work well, models that anticipate are the future of machine learning. Using static relations to predict future movement results in... Read More While machines that analyze movement work well, models that anticipate are the future of machine learning. Using static relations to predict future movement results in less accurate forecasting because relations between static entities change over time. This technology uses dynamic Neural Relational Inference to predict an outcome for each static relation over time, resulting in more accurate forecasts.