Android Security Research
DBank and FARM have both detected two previously unknown Android banking trojans and rooting malware, respectively. These four have been confirmed by the Google Android Security Team.
BEEF
The explanations provided by BEEF are in natural language, and are balanced, meaning that they can contain supporting reasons why the prediction should be true, but also why the prediction might be false.
Boko Haram
Researchers now at NSAIL discovered that not shutting down Boko Haram locations is positively linked to nonoccurrence of sexual violence, arson attacks and suicide bombings.
Bot Research
SentiBot placed first in the DARPA Twitter Bot Challenge against 5 other teams, achieving high accuracy and speed.
Chinese Incursions into India
Chinese incursions across the Indian border have been increasing in recent years. Read more about when China carries out these incursions and where.
Cyber Deception Research
The Stackelberg Honey-based Adversarial Reasoning Engine performs very well, even when the adversary deviates from the initial assumptions made about his or her behavior.
DUCK
Terrorist groups are increasingly using drones to carry out attacks. DUCK is a testbed that shows how to protect cities around the world using a mix of logic, agents, and game theory.
DiscX
DiscX is intended to augment the current decision making procedure for “exploiting vs. disclosing” with a rigorous tool that uses agency experts’ inputs to help agencies such as the US Government’s Equities Review Board arrive at an optimal solution.
FORGE
A project which generates fake technical documents to deceive and impose additional costs on malicious actors.
IM
NSAIL Researchers were the first to build a machine learning based predictive model of any terror group. Our work on understanding Indian Mujahideen (IM) has led to significant policy breakthroughs.
LeT
NSAIL Researchers were the first to build a machine learning based predictive model of any terror group. Our work on understanding LeT has led to significant policy breakthroughs.
NTEWS
We develop methods to forecast the types of terror attacks and the approximate time frames of those attacks for six well known terror groups.
PCORE
NSAIL Researchers were the first to build a machine learning based model to characterize the risk of pastoral conflict in Central Africa.
Phishing
The LNU (Lichtenstein-Northwestern University) Phishing dataset Is the largest dataset appropriate for reproducible research on phishing URLs.
PLATO
NSAIL Researchers discovered that the most significant features for predicting Al Qaeda’s lethality are related to their Public Communications and Logistical subnetworks, while the Leadership and Operational subnetworks are most impactful for predicting ISIS’s lethality.
Sockpuppet Research
We tested the STARS attack on 7 recent review fraud detectors and found that all of them can be severely compromised by the STARS attack on all 4 datasets that we tested — from Amazon, Epinions, Bitcoin Alpha, and Bitcoin OTC exchanges.
STAR
Results show that real world timed association rules have high redundancy and our algorithms are able to summarize them in a very effective manner.
STONE
NSAIL Researchers were the first to develop AI techniques to destabilize a terrorist network and show their efficacy on groups such as Hamas, Hezbollah, Lashkare- Taiba, and Indian Mujahideen.
TREAD
Can deepfakes be used to counter terrorists and destabilize terror groups? The TREAD project shows how to do so.
VEST
VEST is the first system to predict when a vulnerability will be exploited and how severe it will be.
Video Deception
The POLLY Political Lying framework comprises the largest multicultural, multinational, multilingual video dataset that examines political deception.