When neuroscientists David Hubel and Torsten Wiesel wanted to figure out how the brain parses its visual environment, they went as simple as they could go. In a Harvard lab crammed with electrical equipment, they positioned cats in front of a screen and showed them extremely basic images: dots in particular locations, lines at various angles. At the same time, they used implanted electrodes to, quite literally, "listen" to neurons in the areas of the brain devoted to vision. By observing which neurons fired in response to which shapes, they were able to unlock a part of the brain's "visual code," the way in which it represents visual information about its environment. For their achievement, Hubel and Wiesel won the Nobel Prize in 1981, and their discoveries kick-started the rich, diverse field of visual neuroscience.

The production of lithium-ion battery cells is characterized by a high degree of complexity due to numerous cause-effect relationships between process characteristics. Knowledge about the multi-stage production is spread among several experts, rendering tasks as failure analysis challenging. In this paper, a new method is presented that includes expert knowledge acquisition in production ramp-up by combining Failure Mode and Effects Analysis (FMEA) with a Bayesian Network. Special algorithms are presented that help detect and resolve inconsistencies between the expert-provided parameters which are bound to occur when collecting knowledge from several process experts. We show the effectiveness of this holistic method by building up a large scale, cross-process Bayesian Failure Network in lithium-ion battery production and its application for root cause analysis.

In this paper, we introduce proximal gradient temporal difference learning, which provides a principled way of designing and analyzing true stochastic gradient temporal difference learning algorithms. We show how gradient TD (GTD) reinforcement learning methods can be formally derived, not by starting from their original objective functions, as previously attempted, but rather from a primal-dual saddle-point objective function. We also conduct a saddle-point error analysis to obtain finite-sample bounds on their performance. Previous analyses of this class of algorithms use stochastic approximation techniques to prove asymptotic convergence, and do not provide any finite-sample analysis. We also propose an accelerated algorithm, called GTD2-MP, that uses proximal ``mirror maps'' to yield an improved convergence rate. The results of our theoretical analysis imply that the GTD family of algorithms are comparable and may indeed be preferred over existing least squares TD methods for off-policy learning, due to their linear complexity. We provide experimental results showing the improved performance of our accelerated gradient TD methods.

The important peaks related to the physical properties of a lithium ion rechargeable battery were extracted from the measured X ray diffraction spectrum by a convolutional neural network based on the Attention mechanism. Among the deep features, the lattice constant of the cathodic active material was selected as a cell voltage predictor, and the crystallographic behavior of the active anodic and cathodic materials revealed the rate property during the charge discharge states. The machine learning automatically selected the significant peaks from the experimental spectrum. Applying the Attention mechanism with appropriate objective variables in multi task trained models, one can selectively visualize the correlations between interesting physical properties. As the deep features are automatically defined, this approach can adapt to the conditions of various physical experiments.

As summer quickly approaches, some dads are itching to get outside. Even if the number of places we can go has been reduced due to the pandemic, many will spend hours in their backyards tinkering with home projects, training for a nonexistent triathlon and grilling every chance they get. As Father's Day approaches, here are the best gifts for all the DIY-, camping-, grilling- and sport-loving dads in our lives. A good head lamp is an easy to way upgrade Dad's camping kit. We've recommended BioLite head lamps in the past, and the new HeadLamp 200 is a winner too, not to mention quite affordable. This model's USB rechargeable battery makes it more convenient than traditional head lamps because your dad won't have to worry about having a few AAA batteries on hand: Just plug it in and charge it up.

Amazon-owned Ring has unleashed a few upgraded versions of the six-year-old Ring Video Doorbell over the years, including the $230 Video Doorbell 3 Plus that went on sale in April, yet the original $100 Ring Video Doorbell has remained on sale as a budget option. Now comes word that the $100 model is finally getting its own update. Available for pre-order now and slated to ship on June 3, the second-generation Ring Video DoorbellRemove non-product link will keep the original's $100 price tag while adding new features such as improved video resolution, privacy zones, an additional "near" motion zone, and more. Starting with the basics, the revamped Ring Video Doorbell boosts its video resolution to 1080p, versus 720p for the first-generation model, along with "crisper" night vision and "improved" two-way audio quality. The new Video Doorbell features "privacy zones" that let you specify areas in the camera's 155-degree field of view that you don't want recorded or displayed in the Ring App's live view, while an additional "near zone" allows for motion detection in areas that are between five and 15 feet in front of your home.

Researchers have designed a machine learning method that can predict battery health with 10x higher accuracy than current industry standard, which could aid in the development of safer and more reliable batteries for electric vehicles and consumer electronics. The researchers, from Cambridge and Newcastle Universities, have designed a new way to monitor batteries by sending electrical pulses into them and measuring the response. The measurements are then processed by a machine learning algorithm to predict the battery's health and useful lifespan. Their method is non-invasive and is a simple add-on to any existing battery system. The results are reported in the journal Nature Communications.

Researchers have developed a machine learning method that can predict battery health with ten times higher accuracy than current industry standard, which could aid in the development of safer and more reliable batteries for electric vehicles and consumer electronics. The researchers, from Cambridge and Newcastle Universities, have designed a new way to monitor batteries by sending electrical pulses into them and measuring the response. The measurements are then processed by a machine learning algorithm to predict the battery's health and useful lifespan. Their method is non-invasive and is a simple add-on to any existing battery system. The results are reported here.

Metal–air batteries, consisting of a metal anode and an air cathode, have been attracted significant interest by the research community as energy storage devices, because of their high energy density (in particular, compared to lithium ion batteries -LIBs-). A wide diversity of active metals can be used as anode material such as Li, Ca, Mg, Al, Fe, and Zn. However, they have so far found their use only in very particular markets requiring high energy density such as hearing aids. Indeed, despite very significant experimental research efforts, recharging them electrochemically constitute a significant challenge, that if unlocked, will pave the way to a wider diversity of ZAB applications such as Electric Vehicles. Reversing this process to recharge electrochemically a ZAB would imply a heterogeneous deposition of Zn in the anode and the formation of dendrites that can short-circuit the cell, similarly to what can happen in lithium metal batteries.

Researchers at two U.K. universities have developed a way to predict battery health with 10 times greater accuracy than the current industry standard. A machine learning method developed by researchers at the University of Cambridge and Newcastle University in the U.K. can predict battery health with 10 times greater accuracy than the current industry standard. The new method could help develop safer, more reliable batteries for electric vehicles and consumer electronics. The researchers trained the model by performing more than 20,000 experimental measurements. Said Cambridge's Alpha Lee, "By improving the software that monitors charging and discharging, and using data-driven software to control the charging process, I believe we can power a big improvement in battery performance."