Steven Pinker is a cognitive psychologist, linguist, and author of Bill Gates' two favorite books. However, his latest – Enlightenment Now – has some serious shortcomings centering on Pinker's misperceptions about climate change polarization. Pinker falls into the trap of'Both Siderism,' acknowledging the Republican Party's science denial, but also wrongly blaming liberals for the policy stalemate, telling Ezra Klein: There are organizations like Greenpeace and NRDC who are just dead set opposed to nuclear. There are also people on the left like Naomi Klein who are dead set against carbon pricing because it doesn't punish the polluters enough ... the people that you identify who believe in a) carbon pricing and b) expansion of nuclear power, I suspect they're a tiny minority of the people concerned with climate … What we need are polling data on how many people really would support carbon pricing and an expansion of nuclear and other low carbon energy sources. Here Pinker has created a strange straw man that bears no resemblance to the real population of American liberals and environmentalists.

After a satisfying meal of Chinese takeout, you absentmindedly crack open the complimentary fortune cookie. Glancing at the fortune inside, you read, "A dream you have will come true." Scoffing, you toss the small piece of paper and pop the cookie in your mouth. Being the intelligent, well-reasoned person you are, you know the fortune is insignificant--no one can predict the future. However, that thought may be incomplete.

The GraphGrail Ai project has given an interview to Simon Coking, Editor in Chief of IrishTechNews, a leading blockchain technology news resource.

The use of back-propagation and its variants to train deep networks is often problematic for new users, with issues such as exploding gradients, vanishing gradients, and high sensitivity to weight initialization strategies often making networks difficult to train. In this paper, we present Local Representation Alignment (LRA), a training procedure that is much less sensitive to bad initializations, does not require modifications to the network architecture, and can be adapted to networks with highly nonlinear and discrete-valued activation functions. Furthermore, we show that one variation of LRA can start with a null initialization of network weights and still successfully train networks with a wide variety of nonlinearities, including tanh, ReLU-6, softplus, signum and others that are more biologically plausible. Experiments on MNIST and Fashion MNIST validate the performance of the algorithm and show that LRA can train networks robustly and effectively, succeeding even when back-propagation fails and outperforming other alternative learning algorithms, such as target propagation and feedback alignment.

Deploying state-of-the-art CNNs requires power-hungry processors and off-chip memory. This precludes the implementation of CNNs in low-power embedded systems. Recent research shows CNNs sustain extreme quantization, binarizing their weights and intermediate feature maps, thereby saving 8-32 memory and collapsing energy-intensive sum-of-products into XNOR-and-popcount operations. We present XNORBIN, an accelerator for binary CNNs with computation tightly coupled to memory for aggressive data reuse. Implemented in UMC 65nm technology XNORBIN achieves an energy efficiency of 95 TOp/s/W and an area efficiency of 2.0 TOp/s/MGE at 0.8 V. I.

We propose a closed-loop power control algorithm for the downlink of the voice over LTE (VoLTE) radio bearer for an indoor environment served by small cells. The main contributions of our paper are: 1) proposing closed-loop power control for downlink VoLTE (or any packetized voice bearer), 2) deriving an upper bound of the loss in VoLTE downlink signal to noise plus interference ratio which the closed-loop power control has to overcome, 3) employing reinforcement learning to perform closed-loop power control, and 4) showing that this closed-loop power control method can improve the quality of VoLTE in a realistic network setup. Our simulation results have shown that our proposed algorithm significantly improved both voice retainability and mean opinion score as a result of maintaining the effective downlink signal to interference plus noise ratio against adverse network operational issues and faults.

A version of this article appeared in the Autumn 2017 issue of strategy business. But just such a change appears to be happening now. In a great wave of technological change, sensors are spreading through factories and warehouses, software is predicting the need for maintenance before a machine breaks down, power grids and loading docks are becoming intelligent, and custom-designed parts are being produced on demand. The leaders of the next industrial revolution are companies making advances in fields such as robotics, machine learning, digital fabrication (including 3D printing), the Industrial Internet, the Internet of Things (IoT), data analytics and blockchain (a system of decentralized, automated transaction verification). Because these technologies all reinforce the others' impact, they are leading to a new level of proficiency, and to new types of opportunities and challenges for business and for society at large. One key indicator is that conventional boundaries between industries are eroding. It's getting harder to tell the difference between, say, a telecommunications company and an entertainment producer, or between a retail bank and a retail store. The relationships among suppliers, producers, and consumers are also blurring, more rapidly than many business decision makers are prepared for. The foundation of business strategy has long been the classic value chain, which links together raw materials producers, manufacturers, distributors, and (in the end) consumers through a well-established commercial infrastructure characterized by a stable set of transactions. But the rise of digital technology enables individuals to connect outside the value chain and deliver more efficient, effective products and services. This will reduce the importance of economies of scale and conventional divisions of labor.

But consider this: every day, US companies tease apart chemicals in billions of reactions to make food and beverages, drugs, and fuel. In fact, this process is so common in industrial settings that it uses as much energy as all US cars and trucks combined. "It represents 12 percent of all US energy consumption," says Shreya Dave, cofounder of the year-old startup Via Separations. That's largely because separation technology relies largely on a wasteful and time-consuming procedure that's hardly changed in 100 years: using heat to boil and condense chemicals into a pure form. Filtering liquids with a membrane is more efficient, but it's difficult to find one stable enough to avoid reacting with the chemicals it's supposed to filter, or fine enough to filter nanoparticles.

To remain livable and economically competitive, rapidly growing cities need to embrace high-tech solutions to solve their many practical problems. However, how willing are citizens to sacrifice their privacy for the benefits of smart cities, and can government regulations keep up with new tech? This places a significant burden on vital infrastructure, such as transport, housing, energy supply, health care and waste management. Livability in the megacities of tomorrow will largely be determined by the smart solutions being developed today. The term "smart city" is popular among policymakers worldwide.

Computing the partition function $Z$ of a discrete graphical model is a fundamental inference challenge. Since this is computationally intractable, variational approximations are often used in practice. Recently, so-called gauge transformations were used to improve variational lower bounds on $Z$. In this paper, we propose a new gauge-variational approach, termed WMBE-G, which combines gauge transformations with the weighted mini-bucket elimination (WMBE) method. WMBE-G can provide both upper and lower bounds on $Z$, and is easier to optimize than the prior gauge-variational algorithm. We show that WMBE-G strictly improves the earlier WMBE approximation for symmetric models including Ising models with no magnetic field. Our experimental results demonstrate the effectiveness of WMBE-G even for generic, nonsymmetric models.