Examples include early diagnosis of medical diseases [22], threat detection for cyber-attacks [3, 18, 36], or safety analysis for self-driving cars [38]. Manyreal-world anomalies can be detected intime series data.

Undetected anomalies in time series can trigger catastrophic failures in safety-critical systems, such as chemical plant explosions or power grid outages. Although many detection methods have been proposed, their performance remains unclear because current metrics capture only narrow aspects of the task and often yield misleading results. We address this issue by introducing verifiable properties that formalize essential requirements for evaluating time-series anomaly detection. These properties enable a theoretical framework that supports principled evaluations and reliable comparisons. Analyzing 37 widely used metrics, we show that most satisfy only a few properties, and none satisfy all, explaining persistent inconsistencies in prior results. To close this gap, we propose LARM, a flexible metric that provably satisfies all properties, and extend it to ALARM, an advanced variant meeting stricter requirements.

In this paper, we provide a systematic approach for assessing and comparing the computational complexity of neural network layers in digital signal processing. We provide and link four software-to-hardware complexity measures, defining how the different complexity metrics relate to the layers' hyper-parameters. This paper explains how to compute these four metrics for feed-forward and recurrent layers, and defines in which case we ought to use a particular metric depending on whether we characterize a more soft- or hardware-oriented application. One of the four metrics, called `the number of additions and bit shifts (NABS)', is newly introduced for heterogeneous quantization. NABS characterizes the impact of not only the bitwidth used in the operation but also the type of quantization used in the arithmetical operations. We intend this work to serve as a baseline for the different levels (purposes) of complexity estimation related to the neural networks' application in real-time digital signal processing, aiming at unifying the computational complexity estimation.

If you need a little help cleaning up after the holidays, you've come to the right place. Amazon's selling the iRobot Roomba 694 robot vacuum for just $159, which is 42 percent off of the original $274.99 price. This device comes with an auto-adjust cleaning head, which changes the height of the brushes depending on the floor type, and has a runtime of approximately 90 minutes. According to Amazon reviews, the quality of the brush rolls are truly top-notch, which really makes a difference when grabbing debris. The iRobot Roomba 694 offers customized schedules and is capable of cleaning both carpets and hardwood floors.

Deep neural networks (DNNs) are vulnerable to backdoor attack, which does not affect the network's performance on clean data but would manipulate the network behavior once a trigger pattern is added. Existing defense methods have greatly reduced attack success rate, but their prediction accuracy on clean data still lags behind a clean model by a large margin. Inspired by the stealthiness and effectiveness of backdoor attack, we propose a simple but highly effective defense framework which injects non-adversarial backdoors targeting poisoned samples. Following the general steps in backdoor attack, we detect a small set of suspected samples and then apply a poisoning strategy to them. The non-adversarial backdoor, once triggered, suppresses the attacker's backdoor on poisoned data, but has limited influence on clean data. The defense can be carried out during data preprocessing, without any modification to the standard end-to-end training pipeline. We conduct extensive experiments on multiple benchmarks with different architectures and representative attacks. Results demonstrate that our method achieves state-of-the-art defense effectiveness with by far the lowest performance drop on clean data. Considering the surprising defense ability displayed by our framework, we call for more attention to utilizing backdoor for backdoor defense. Code is available at https://github.com/damianliumin/non-adversarial_backdoor.

Did you miss a session at the Data Summit? During the pandemic, hiring has become increasingly remote as companies eschew job fairs and other in-person recruitment initiatives in favor of alternatives. According to a LinkedIn survey, companies plan to adopt hiring processes that combine virtual and in-person steps due to the associated cost and time savings. But priorities in HR largely haven't changed. Jobvite's 2021 poll found that improving quality-of-hire, improving time-to-hire, increasing the retention rate, and growing the talent pipeline remain recruiters' top recruiting priorities.

Early in the pandemic, as many enterprises looked to digitally transform their operations, marketing and merchandising departments turned to AI to automate the increasing workloads. According to Salesforce, marketers' use of AI soared between 2018 and 2020, jumping from 29% in 2018 to 84% in 2020. A separate survey from ManageEngine -- the IT division of Zoho -- found that analytics for marketing, driven by automation and AI, experienced a 44% adoption surge in over the past two years. Retail and consumer products executives, responding to a recent IBM survey, said that they believe that intelligent automation capabilities could help increase annual revenue growth by up to 10%. Adopters in marketing cite benefits like accelerating revenue as well as getting actionable insights from marketing data.

Placer.ai, a location analytics platform that serves companies with data around consumer foot traffic, has raised $100 million in a series C round of funding, valuing the company at $1 billion. The location intelligence industry was pegged as a $12 billion market last year, a figure that's predicted to more than double in the coming years as businesses leverage big data insights to improve their bottom line. For example, businesses can glean accurate foot traffic counts and "dwell time," allowing them to filter by time and day as well as customer segments. This can be useful for understanding how special promotions, events, or holidays impact trade. Or they can discover what Placer.ai

After playing through some Billie Eilish tracks in Beat Saber, soon you'll also be able to kick back and listen to a limited-edition Echo Studio sporting the cover of her latest album, "Happier Than Ever." Beyond the beige fabric and Eilish's visage, the $230 speaker is no different than the standard $200 Echo Studio. That's a shame if you were hoping for some sort of upgrade, but if it's any consolation, we adored the Echo Studio's beefy hardware when it launched two years ago. It's one of the few smart speakers built for 3D Audio, and it has more than enough power to blast all of your favorite tunes. The Billie Eilish Limited-Edition Echo Studio also comes with a six-month subscription to Amazon Music, typically a $48 value.

Fetcher.ai, a recruitment platform that combines AI with human teams, today announced it has raised $6.5 million in a round led by G20 Ventures. The company, whose latest funding round brings its total to $12 million, says the funds will be used to expand the size of its workforce. In 2020, talent shortages in the U.S. rose to historic levels, with 69% of employers reporting having difficulty filling jobs, according to a ManPowerGroup survey. A report by the Society for Human Resource Management found that filling an open position costs employers an average of $4,129 and takes roughly 42 days. Fetcher was founded in 2014 as a consumer-centric messaging app called Caliber, which focused on professional networking.