Optimized Disk Imaging and Forensics in Java && Using Deep Learning in Java
Talk #1: Optimized Disk Imaging and Forensics in Java
Tupelo is an open-source Java codebase for systematic whole disk acquisition, storage and search. The system is designed to handle many disks being imaged many times, perhaps once a week or even daily. The acquisition and storage phases use some clever tricks to optimize both time and space requirements. Using trusted boot media, disk contents can be imaged without reliance on potentially infected system software. Use of FUSE enables fully searchable stored disks, allowing traversals by e.g. Sleuthkit for accurate volume system and file system layout.
Tupelo builds upon various useful Java libraries: virtual machine disk access from the host, split Java/C artifact construction for JNI projects, a Windows Registry reader, a STIX engine based on JAXB, and a Java wrapper to the open-source Sleuthkit disk forensics toolset.
Speaker: Stuart Maclean
Stuart Maclean is an engineer at the University of Washington. He has built cybersecurity tools to answer one simple question: What does this program do? One obvious thing it might do is store data on your disk, perhaps in a filesystem, perhaps not. A method for whole disk capture and later search is therefore needed, and the toolset described in this talk goes some way to answering such questions.
Talk #2: Using Deep Learning in Java
Being tired of having to manually label all the photos he takes on weekend outdoor trips, Eric decided to try to get Deeplearning4j [ https://deeplearning4j.org/ ] to do the job for him...
Speaker: Eric Jain
http://www.nimret.org/seajug/past/2017/oct/eric.jpg Eric Jain [ https://www.linkedin.com/in/ejain/ ] runs Zenobase [ https://zenobase.com/ ], a service that helps people answer questions using data from their fitness-tracking devices (and other sources).