We adopt a deep learning method casi (Convolutional Approach to Shell Identification) and extend it to 3D (casi-3d) to identify signatures of stellar feedback in molecular line spectra, such as 13 CO. We adopt magneto-hydrodynamics simulations that study the impact of stellar winds in a turbulent molecular cloud as an input to generate synthetic observations. We apply the 3D radiation transfer code radmc-3d to model 13 CO (J=1-0) line emission from the simulated clouds. We train two casi-3d models: ME1 is trained to predict only the position of feedback, while MF is trained to predict the fraction of the mass coming from feedback in each voxel. We adopt 75% of the synthetic observations as the training set and assess the accuracy of the two models with the remaining data. We demonstrate that model ME1 identifies bubbles in simulated data with 95% accuracy, and model MF predicts the bubble mass within 4% of the true value. We test the two models on bubbles that were previously visually identified in Taurus in 13 CO. We show our models perform well on the highest confidence bubbles that have a clear ring morphology and contain one or more sources. We apply our two models on the full 98 deg 2 FCRAO 13 CO survey of the Taurus cloud. Models ME1 and MF predict feedback gas mass of 2894 M and 302 M , respectively. When including a correction factor for missing energy due to the limited velocity range of the 13 CO data cube, model ME1 predicts feedback kinetic energies of 4.0×10 46 ergs and 1.5×10 47 ergs with/without subtracting the cloud velocity gradient. Model MF predicts feedback kinetic energy of 9.6×10 45 ergs and 2.8×10 46 ergs with/without subtracting the cloud velocity gradient. Model ME1 predicts bubble locations and properties consistent with previous visually identified bubbles. However, model MF demonstrates that feedback properties computed based on visual identifications are significantly over-estimated due to line of sight confusion and contamination from background and foreground gas.

We adopt the deep learning method casi-3d (Convolutional Approach to Structure Identification-3D) to identify protostellar outflows in molecular line spectra. We conduct magneto-hydrodynamics simulations that model forming stars that launch protostellar outflows and use these to generate synthetic observations. We apply the 3D radiation transfer code radmc-3d to model 12 CO (J=1-0) line emission from the simulated clouds. We train two casi-3d models: ME1 is trained to predict only the position of outflows, while MF is trained to predict the fraction of the mass coming from outflows in each voxel. The two models successfully identify all 60 previously visually identified outflows in Perseus. Additionally, casi-3d finds 20 new high-confidence outflows. All of these have coherent high-velocity structure, and 17 of them have nearby young stellar objects, while the remaining three are outside the Spitzer survey coverage. The mass, momentum and energy of individual outflows in Perseus predicted by model MF is comparable to the previous estimations. This similarity is due to a cancelation in errors: previous calculations missed outflow material with velocities comparable to the cloud velocity, however, they compensate for this by over-estimating the amount of mass at higher velocities that has contamination from non-outflow gas. We show outflows likely driven by older sources have more high-velocity gas compared to those driven by younger sources.

A large and realistic dataset to support research in the field of object geolocalization• An object detector designed to predict GPS locations using a local offset and coordinate transform• A tracker which condenses object detections into geolocalized predictions

Host-based Intrusion Detection Systems (HIDS) automatically detect events that indicate compromise by adversarial applications. HIDS are generally formulated as analyses of sequences of system events such as bash commands or system calls. Anomaly-based approaches to HIDS leverage models of normal (aka baseline) system behavior to detect and report abnormal events, and have the advantage of being able to detect novel attacks. In this paper we develop a new method for anomaly-based HIDS using deep learning predictions of sequence-to-sequence behavior in system calls. Our proposed method, called the ALAD algorithm, aggregates predictions at the application level to detect anomalies. We investigate the use of several deep learning architectures, including WaveNet and several recurrent networks. We show that ALAD empowered with deep learning significantly outperforms previous approaches. We train and evaluate our models using an existing dataset, ADFA-LD, and a new dataset of our own construction, PLAID. As deep learning models are black box in nature we use an alternate approach, allotaxonographs, to characterize and understand differences in baseline vs.~attack sequences in HIDS datasets such as PLAID.