With increasing online fraud and identity theft each day, service providers need a way to ensure that their services cannot be compromised. Anti-spoofing liveness detection is required especially in unsupervised authentication situations. Biometric authentication systems need to prevent sophisticated spoofing challenges from replay attacks and determine the user's presence. Thus, BioID's presentation attack detection (PAD) is crucial for eKYC onboarding, online login and banking transactions. BioID is a pioneer and the leading player in face liveness detection for assured user presence.
Cowboy is bringing crash detection smarts to its e-bikes by the end of this month. It has been beta testing the feature with a thousand users, but it'll soon be available for free to all Cowboy 2 and Cowboy 3 owners. The bike uses its sensors to watch out for potential falls. If it detects one, and you don't confirm that everything's okay within a minute, Cowboy can alert up to two emergency contacts. If you don't have your phone with you or the bike can't communicate with the Cowboy app, it can use its built-in SIM card to let your contacts know something might be wrong. It'll also provide your location to them in real-time through GPS tracking.
Apple has made another acquisition in the artificial intelligence industry. GeekWire reports today that Apple has acquired Xnor․ai, a Seattle-based startup that focuses on low-power artificial intelligence technology. Xnor․ai notably provided the AI behind smart camera company Wyze's on-device people detection feature; the acquisition by Apple explains why Wyze recently lost that feature: Person Detection will be temporarily removed from Wyze Cam beginning with a new firmware release planned for mid-January 2020, due to the unexpected termination of our agreement with our AI provider. We are preparing to roll out our own in-house version of Person Detection this year, which will remain free for our users. Neither Apple nor Xnor․ai have confirmed the deal, but today's report suggests that Apple paid somewhere in the range of $200 million to complete the acquisition.
When working with network datasets, the theoretical framework of detection theory for Euclidean vector spaces no longer applies. Nevertheless, it is desirable to determine the detectability of small, anomalous graphs embedded into background networks with known statistical properties. Casting the problem of subgraph detection in a signal processing context, this article provides a framework and empirical results that elucidate a detection theory" for graph-valued data. Its focus is the detection of anomalies in unweighted, undirected graphs through L1 properties of the eigenvectors of the graph's so-called modularity matrix. This metric is observed to have relatively low variance for certain categories of randomly-generated graphs, and to reveal the presence of an anomalous subgraph with reasonable reliability when the anomaly is not well-correlated with stronger portions of the background graph. An analysis of subgraphs in real network datasets confirms the efficacy of this approach."
Model change detection is studied, in which there are two sets of samples that are independently and identically distributed (i.i.d.) according to a pre-change probabilistic model with parameter $\theta$, and a post-change model with parameter $\theta'$, respectively. The goal is to detect whether the change in the model is significant, i.e., whether the difference between the pre-change parameter and the post-change parameter $\|\theta-\theta'\|_2$ is larger than a pre-determined threshold $\rho$. The problem is considered in a Neyman-Pearson setting, where the goal is to maximize the probability of detection under a false alarm constraint. Since the generalized likelihood ratio test (GLRT) is difficult to compute in this problem, we construct an empirical difference test (EDT), which approximates the GLRT and has low computational complexity. Moreover, we provide an approximation method to set the threshold of the EDT to meet the false alarm constraint. Experiments with linear regression and logistic regression are conducted to validate the proposed algorithms.