|
Samsung mass producting 128 Gigabit NAND memory using 10 nanometer lithography
Samsung has begun mass producing a 128-gigabit (Gb), 3-bit multi-level-cell (MLC) NAND memory chip using 10 nanometer (nm)-class process technology this month. This chip will enable high-density memory solutions such as embedded NAND storage and solid state drives (SSDs).
"By introducing next-generation memory storage products like the 128Gb NAND chip, Samsung is extremely well situated to meet growing global customer needs," said Young-Hyun Jun, executive vice president, memory sales & marketing, Device Solutions Division, Samsung Electronics. "The new chip is a critical product in the evolution of NAND flash, one whose timely production will enable us to increase our competitiveness in the high density memory storage market."
Samsung's 128Gb NAND flash is based on a 3-bit multi-level-cell design and 10nm-class process technology, which means a process technology node somewhere between 10 and 20 nanometers. It boasts the industry's highest density as well as the highest performance level of 400 megabits-per-second (mbps) data transfer rate based on the toggle DDR 2.0 interface. Read more » Environmentalists are Hypocritical about China's Nuclear, Wind and Solar Power
China is building the most new nuclear power of any country and should have about 40 gigawatts of nuclear power generating about 260 TWh. China's current 15 GW of nuclear power generates about 98 TWh.
China increased its solar PV target to 40 GW.
An MIT researcher estimates that 40 GW of solar PV power in China will generate about 50 TWh. Some environmentalists will try to spin that China's nuclear power does not count because it is centrally planned. Then they say that China's solar power and wind power mean that solar power and wind have arrived even though solar and wind also are part of China's central energy plan. The reality is that China is the single most important energy market in the world. The developing countries are where all of the economic growth is happening and where the vast majority of new energy generation is getting built. China counts for nuclear. China counts for wind and solar. So far, China has added more hydro power (like the Three Gorges dam)  The 40 GW of solar power is cheered by Cleantechnica and is proof that solar power has arrived. Amory Lovins is a considered an energy expert by green environmentalists. In 2008 Amory Lovins said - In 2006 distributed renewables alone got $56 billion of private risk capital while nuclear as usual got zero—it's only bought by central planners. Nuclear added less capacity than photovoltaics and a 10th of what wind power added. Even in China, which has ambitious nuclear goals, they already have seven times as much distributed renewable as nuclear capacity, and it's growing seven times faster. NOTE- Lovins combines multiple lies (or more politely confusing spin). He classifies smaller natural gas power units as distributed micropower. Below you can see that nuclear power generation is about equal to wind and solar. So the growing seven times faster was kw and not kwh and some additional playing with numbers and definitions that do not represent reality. China's power generation in 2012 China's generating capacity in 2012 was 4977.4 TWh, an increase of 5.2% Hydro generating 864.1 TWh, an increase of 29.3%; Thermal power (mainly coal) 3910.8 TWh up by 0.3% Nuclear power 98.2 TWh, up by 12.6% Wind power 100.4 TWh, up 35.5% Solar power generation 3.5 TWh, up by 414%. 3676 kwh / year per capita power generation for about 40% in developed countries. Read more » Google, Baidu, Dwave Systems focused on Sparse Coding for more accurate image classification and unsupervised feature learning
Sparse coding is a hot area in the field of Deep Learning. Deep learning and sparse coding appears highly promising for increasing the accuracy of image classification and it enables a system to look at tens of millions of images and to classify the information without human supervision. Increasing the accuracy of image and voice recognition can transform the interaction and experience that people have with search systems and artificial intelligence interfaces. A highly accurate voice recognition system means that people can just talk to a computer and not need keyboards and mice. This enables efficient and transformative new form factors for devices. Google, Baidu (dominant search engine in China) and Dwave Systems (adiabatic quantum computers) are all focused on Sparse coding and deep learning.
Dwave systems has a 512 qubit adiabatic quantum computer. They recently wrote up how to solve sparse coding using the Dwave system. The results are comparable to the best conventional systems. If Dwave systems are scaled to 2000 qubits the speed gain expected would be about 500,000 times. This would suggest that Dwave would be the fastest systems for important aspects of artificial intelligence and machine learning.
Dwave used the PiCloud python libraries, which allows us to run hundreds or thousands of parallel jobs to perform the optimization over the weights. As a rough estimate, for the optimization problems generated by MNIST, each optimization using FSS takes about 30 milliseconds, and there are 60,000 of these per iteration of the block descent procedure. If we run serially this is about 30 minutes per iteration. If we use 100 cores, we can send 600 jobs to each core, and get about 100x speed-up, taking the time down to about 20 seconds. As an interesting aside, Dwave find that our own Python implementation of FSS is about the same in terms of performance as the original MATLAB code provided by Honglak Lee. This was a little surprising as the core computations run in highly optimized compiled code inside MATLAB. This is evidence that the routines within numpy are competitive with MATLAB's versions for the core FSS computations. Deep Learning Deep learning is a sub-field of machine learning (artificial intelligence) that is based on learning several levels of representations, corresponding to a hierarchy of features or factors or concepts, where higher-level concepts are defined from lower-level ones, and the same lower-level concepts can help to define many higher-level concepts. Deep learning is part of a broader family of machine learning methods based on learning representations. An observation (e.g., an image) can be represented in many ways (e.g., a vector of pixels), but some representations make it easier to learn tasks of interest (e.g., is this the image of a human face?) from examples, and research in this area attempts to define what makes better representations and how to learn them. Sparse Coding The sparse code is a kind of neural code in which each item is encoded by the strong activation of a relatively small set of neurons. For each item to be encoded, this is a different subset of all available neurons. As a consequence, sparseness may be focused on temporal sparseness ("a relatively small number of time periods are active") or on the sparseness in an activated population of neurons. In this latter case, this may be defined in one time period as the number of activated neurons relative to the total number of neurons in the population. This seems to be a hallmark of neural computations since compared to traditional computers, information is massively distributed across neurons. A major result in neural coding from Olshausen et al. is that sparse coding of natural images produces wavelet-like oriented filters that resemble the receptive fields of simple cells in the visual cortex.  Kai Yu, who leads Baidu's Artificial Intelligence Lab, has a tutorial on Deep Learning and Sparse Coding. (69 pages) Read more » Baidu open Deep Learning Artificial Intelligence Lab in Silicon Valley
In late January, word arrived that the Chinese search giant was setting up a research lab dedicated to "deep learning" — an emerging computer science field that seeks to mimic the human brain with hardware and software — and as it turns out, this lab includes an operation here in Silicon Valley, not far from Apple headquarters, in addition to a facility back in China. The company just hired its first researcher in Cupertino, with plans to bring in several more by the end of the year.
Baidu calls its lab The Institute of Deep Learning, or IDL. Much like Google and Apple and others, the company is exploring computer systems that can learn in much the same way people do. "We have a really big dream of using deep learning to simulate the functionality, the power, the intelligence of the human brain," says Kai Yu, who leads Baidu's speech- and image-recognition search team and just recently made the trip to Cupertino to hire that first researcher. "We are making progress day by day." Kai Yu, deputy engineering director of Baidu, has a webpage and a list of publications. In the eyes of CEO Robin Li, Baidu IDL's goal is to become the "AT & T-Bell labs, Xerox PARC this first-class research institutions". If, as Robin Li depicted, this will be the new "Microsoft Asia NARL an important impact on the field of the smart technology. Baidu overall community will benefit from Deep Learning technology advancement with improved search and interface infrastructure. Read more » Predictions for a manned landing on Mars by 2021-2033 and a small permanent Mars Base by 2023-2037
Yesterday I predicted that there would be a permanent base established on Mars by 2037.
This is a prediction where I pad the date to allow for various real life problems to slow down the development. I believe the manned fly by of Mars will happen in 2018. Robert Zubrin is a highly opinionated and outspoken Mars advocate. A few years ago Robert Zubrin proposed going to Mars by 2016. His proposal was mostly technically feasible (although it was counting on using a Spacex Heavy which will not have its first flight until later this year. Also, the Spacex Heavy will not be man rated for a few years.) No group who actually do what Zubrin proposed has stepped up to it. Zubrin was not even predicting it would be done. Zubrin was basically making the technical argument that it should be done and we should accept risk in adventurous space missions. Elon Musk's estimate for a manned mission to Mars is 2024-2027 In 2012, Elon Musk told "Nightline" in an interview at SpaceX headquarters in Los Angeles. "I think we'll be able to send, probably, the first people to Mars in roughly 12 to 15 years. That's my estimate." Note he said that Spacex would be able to send people to Mars around 2024-2027. This is different to saying that they would send people to Mars at that time. He is saying the capability would be there. If Dennis Tito is successful with the Mars flyby in 2018 then it will make it more than 60% likely there would be a human landing on Mars in 2025-2030. However, if the 2018 window is missed the next free return from Mars would be 2031. Therefore, I predicted a Mars colony by 2037. Current technology takes about 250 days to go one way to Mars. The capability will be there by 2027. The manned landing would be by 2024-2033. The base would be by 2028-2035. Mars One - Could be successful but only if reusable Spacex Rocket is developed to bring the cost down Mars One plans to establish the first human settlement on Mars by April 2023. The first crew of four astronauts emigrate to their new planet from Earth, a journey that takes seven months. A new team will join the settlement every two years. By 2033 there will be over twenty people living, working and flourishing on Mars, their new home. The business plan is to use reality TV and other revenue to support the costs of the project. Mars One has a bold plan, but I think they can only afford it if Spacex is successful with the reusable rocket to lower costs for going to space by 100 times.  Read more » Color Printing at 100,000 dots per inch at the diffraction limit of visible light
Commercial laser printers typically produce pin-sharp images with spots of ink about 20 micrometers apart, resulting in a resolution of 1,200 dots per inch (dpi). By shrinking the separation to just 250 nanometers — roughly 100 times smaller — a research team at A*STAR can now print images at an incredible 100,000 dpi, the highest possible resolution for a color image. These images could be used as minuscule anti-counterfeit tags or to encode high-density data.
To print the image, the team coated a silicon wafer with insulating hydrogen silsesquioxane and then removed part of that layer to leave behind a series of upright posts of about 95 nanometers high. They capped these nanoposts with layers of chromium, silver and gold (1, 15 and 5 nanometers thick, respectively), and also coated the wafer with metal to act as a backreflector.  Nature Nanotechnology - Printing colour at the optical diffraction limit Read more » What will happen far sooner than other's have recently predicted
The BBC has collected predictions (from 2013 to 2150) and produced odds for their occurrence. Odds were produced by Ladsbrokes a bet maker in the UK.
The BBC predictions are from Jan 2, 2013. BBC - A successful demonstration of fusion power by 2020. NBF more power will be produced in a nuclear fusion demonstration than the power input before Dec 31, 2017. NBF tracks fusion energy projects all the time. There are some decent prospects for a breakthrough by Dec 31, 2015. The Lawrenceville plasma physics project, John Slough's tests this summer for fusion propulsion could also be a breakthrough net gain lab demonstration of fusion power. General Fusion in Canada could have a good demo within 3 years. EMC2 fusion and Tri-alpha Energy's work seems to be slower but they are both being secretive. BBC - Wealthy people are able to select elements of their offsprings genetic makeup by 2050. NBF - Selection based on genetic makeup will occur by 2023 for pre-implantation and by 2030 for selection of the egg or sperm via non-destructive methods and genetic modification of cells and embroys. Embryos created with assisted reproduction techniques can already have pre-implantation genetic diagnosis and genetic profiling is becoming more advanced. The genes that have a positive or negative effect on intelligence will be announced in two months. 750 genes that have an effect on height are known. Full genome sequencing of an embyro, combined with the knowledge of which genes contribute would allow Read more » More Recent Articles
| ||||||||||||||||||||||
| Your requested content delivery powered by FeedBlitz, LLC, 9 Thoreau Way, Sudbury, MA 01776, USA. +1.978.776.9498 |
