No abstract

Machine learning models are prone to making incorrect predictions on inputs that are far from the training distribution. This hinders their deployment in safety-critical applications such as autonomous vehicles and healthcare. The detection of a shift from the training distribution of individual datapoints has gained attention. A number of techniques have been proposed for such out-of-distribution (OOD) detection. But in many applications, the inputs to a machine learning model form a temporal sequence. Existing techniques for OOD detection in time-series data either do not exploit temporal relationships in the sequence or do not provide any guarantees on detection. We propose using deviation from the in-distribution temporal equivariance as the non-conformity measure in conformal anomaly detection framework for OOD detection in time-series data. Computing independent predictions from multiple conformal detectors based on the proposed measure and combining these predictions by Fisher's method leads to the proposed detector CODiT with guarantees on false detection in time-series data. We illustrate the efficacy of CODiT by achieving stateof-the-art results on computer vision datasets in autonomous driving. We also show that CODiT can be used for OOD detection in non-vision datasets by performing experiments on the physiological GAIT sensory dataset. Code, data, and trained models are available at https://github.com/kaustubhsridhar/time-series-OOD.

No abstract

Deep neural network (DNN) models have proven to be vulnerable to adversarial attacks. In this paper, we propose VisionGuard, a novel attack-and dataset-agnostic and computationally-light defense mechanism for adversarial inputs to DNN-based perception systems. In particular, VisionGuard relies on the observation that adversarial images are sensitive to lossy compression transformations. Specifically, to determine if an image is adversarial, VisionGuard checks if the output of the target classifier on a given input image changes significantly after feeding it a transformed version of the image under investigation. Moreover, we show that VisionGuard is computationally-light both at runtime and design-time which makes it suitable for real-time applications that may also involve large-scale image domains. To highlight this, we demonstrate the efficiency of VisionGuard on ImageNet, a task that is computationally challenging for the majority of relevant defenses. Finally, we include extensive comparative experiments on the MNIST, CIFAR10, and ImageNet datasets that show that VisionGuard outperforms existing defenses in terms of scalability and detection performance.

No abstract

No abstract

-SONOS memory cell using a split gate structure is studied using simulations. The dependence of channel hot electron (HE) IntroductionSplit gate memory cells using source-side-injection (SSI) for programming have generated much interest recently [1][2]. SSI gives higher injection efficiency, faster programming speed and lower power consumption compared to channel hot electron (CHE) injection [1][2][3]. Most of the split gate cells studied so far use floating gate for charge storage. Using nitride for charge storage gives the additional benefits of SONOS type memories [5][6]. SONOS memories are usually programmed using CHE injection and erased using band-to-band tunnelling induced hot hole erase (HHE) for low power and multi bit operation. However, in these, the mismatch in spatial distribution of the trapped electrons and holes is a severe problem affecting the device endurance and retention [4,[8][9]. In this paper, we show through simulations that the hot electron (HE) and hot hole (HH) injection points (during program and erase, respectively) can be independently controlled by using a split gate structure. We first study the effect of trapped charge position on IV (I D -V PG ) characteristics, followed by the effect of program and erase biases on channel HE and HH profiles, respectively. Finally, the effect of program gate length and channel doping on the hot carrier profiles is studied.

Environments with sparse rewards and long horizons pose a significant challenge for current reinforcement learning algorithms. A key feature enabling humans to learn challenging control tasks is that they often receive expert intervention that enables them to understand the high-level structure of the task before mastering low-level control actions. We propose a framework for leveraging expert intervention to solve long-horizon reinforcement learning tasks. We consider option templates, which are specifications encoding a potential option that can be trained using reinforcement learning. We formulate expert intervention as allowing the agent to execute option templates before learning an implementation. This enables them to use an option, before committing costly resources to learning it. We evaluate our approach on three challenging reinforcement learning problems, showing that it outperforms state-of-the-art approaches by an order of magnitude. Project website: https://sites.google.com/view/stickymittens In order to apply RL effectively, to practical applications with a high-dimensional state and action space, exploration is a challenge. Intuitively, in such settings, complex sequences of actions are required to achieve any nonzero reward, which means that random exploration will take extremely long to find a nonzero reward signal. Thus, learning to improve performance can be very slow.Options are an RL tool to circumvent this problem (Sutton et al., 1999). Options are policies designed to achieve intermediate subgoals. For instance, in robot grasping tasks, an option might enable the robot to grasp a block, which