Dealing with the memory problem in artificial intelligence

admin pypqskzmbfie , , , , , , ,

first_imgArtificial Intelligence (AI), which allows computers to learn and perform specific tasks such as image recognition and natural language processing, is loosely inspired by the human brain.The challenge is that while the human brain has evolved over the last 3 million years, Artificial Neural Networks, the very “brain” of AI, which have only been around for a few decades and aren’t nearly as finely tuned or as sophisticated as the gray matter in our heads, are expected to perform tasks associated with human intelligence. So in our quest to create AI systems that can benefit society — in areas from image classification to voice recognition and autonomous driving — we need to find new paths to speed up the evolution of AI.Part of this process is figuring out what types of memory work best for specific AI functions and discovering the best ways to integrate various memory solutions together. From this standpoint, AI faces two main memory limitations: density and power efficiency. AI’s need for power makes it difficult to scale AI outside of datacenters where power is readily available — particularly to the edge of the cloud where AI applications have the highest potential and value.To enable AI at the edge, developments are being made toward domain-specific architectures that facilitate energy-efficient hardware. However, the area that will open the way for the most dramatic improvements, and where a large amount of effort should be concentrated, is the memory technology itself.Driving AI to the Edge Over the last half-century, a combination of public and private interest has fueled the emergence of AI and, most recently, the advent of Deep Learning (DL). DL models have — due to the exceptional perception capabilities they offer — become one of the most widespread forms of AI. A typical DL model must first be trained on massive datasets (typically on GPU servers in datacenters) to tune the network parameters, an expensive and lengthy process, before it can be deployed to make its own inferences based on input data (from sensors, cameras, etc.). DL models require such massive amounts of memory to train their many parameters that it becomes necessary to utilize off-chip memory. Hence, much of the energy cost during training is incurred because of the inefficient shuffling of gigantic data loads between off-chip DRAM and on-chip SRAM (an approach that often exceeds 50% of the total energy use). Once the model is trained, the trained network parameters must be made available to perform inference tasks in other environments.Until recently AI applications had been confined to datacenters because of their large energy consumption and space requirements. However, over the past few years the growing demand for AI models at high scale, low latency and low cost has been pushing these applications to be run at the edge, namely in IoT and on mobile devices where power and performance are highly constrained.This is driving a rapidly expanding hardware ecosystem that supports edge applications for inference tasks and even a nascent effort at enabling distributed training (e.g., Google’s Federated Learning models). These new architectures are primarily driven by speech recognition and image classification applications.The growing demand combined with the increasing complexity of DL models is unsustainable in that it is causing a widening gap between what companies require in terms of energy dissipation, latency and size, and what current memory is capable of achieving. With the end of Moore’s Law and Dennard Scaling in the rear-view mirror exacerbating the issue, the semiconductor industry needs to diversify toward new memory technologies to address this paradigm shift and fulfill the demand for cheap, efficient AI hardware.The Opportunity for New Memories The AI landscape is a fertile ground for innovative memories with unique and improving characteristics and presents opportunities in both the datacenter and at the edge. New memory technologies can meet the demand for memory that will allow edge devices to perform DL tasks locally by both increasing the memory density and improving data access patterns, so that the need for transferring data to and from the cloud is minimized. The ability to perform perception tasks locally, with high accuracy and energy efficiency is key to the further advancement of AI.This realization has led to significant investment in alternative memory technologies, including NAND flash, 3D XPoint (Intel’s Optane), Phase-Change Memory (PCM), Resistive Memory (ReRAM), Magneto-Resistive Memory (MRAM) and others that offer benefits such as energy efficiency, endurance and non-volatility. While facilitating AI at the edge, such memories may also allow cloud environments to perform DL model training and inference more efficiently. Additional benefits include the potential improvements in reliability and processing speed. These improvements in the memory technology will make it possible to circumvent the current hardware limitations of devices at the edge.In particular, certain new memories offer distinct benefits due to specific inherent or unique qualities of the technology for a number of AI applications. ReRAM and PCM offer advantages for inference applications due to their superior speed (compared to Flash), density and non-volatility. MRAM offers similar advantages to ReRAM and PCM; furthermore, it exhibits ultra-high endurance such that it can compete with and complement SRAM as well as function as Flash replacement. Even at these early stages of their lifetime, these new memory technologies show enormous potential in the field of AI.And although we are still decades away from implementing the AI we’ve been promised in science fiction, we are presently on the cusp of significant breakthroughs that will affect many aspects of our lives and provide new efficient business models. As Rockwell Anyoha writes in a Harvard special edition blog on AI, “In the first half of the 20th century, science fiction familiarized the world with the concept of artificially intelligent robots. It began with the ‘heartless’ Tin Man from the Wizard of Oz and continued with the humanoid robot that impersonated Maria in Metropolis.”The next competitive battle is being fought in memory, and as a result, there is a tremendous amount of time, money and brain power being dedicated to figuring out how to fix AI’s memory problem. Ultimately, while these computerized brains don’t yet hold a candle to our human brains — especially pertaining to energy efficiency — it is that very uniqueness of our own minds that enables our capacity to create solutions to our many fantasies and bring artificial intelligence to life.— Andrew Walker is vice president of product at Spin Memory (formerly Spin Transfer Technologies). >> This article was originally published on our sister site, EE Times:”AI’s Memory Problem.” Share this:TwitterFacebookLinkedInMoreRedditTumblrPinterestWhatsAppSkypePocketTelegram Tags: Advanced Technology Continue Reading Previous Seven design challenges for ultrasound smart probesNext A logical method of debugging embedded systemscenter_img Leave a Reply Cancel reply You must Register or Login to post a comment. This site uses Akismet to reduce spam. Learn how your comment data is processed.last_img

You May Also Like..

Start Investing in People

first_img Essential Reading! Get my first book: The Only Sale Guide You’ll Ever Need “The USA Today bestseller by the star sales speaker and author of The Sales Blog that reveals how all salespeople can attain huge sales success through strategies backed by extensive research and experience.” Buy Now We try to help people, but we aren’t giving them the help they really need. It’s because we are underinvesting in people. Investing more means caring more.We give people answers when we should be teaching them to think. But providing an answer takes only a few minutes and teaching people to think takes more time and real effort.We give people money when we should be teaching them to create increasingly greater levels of value. Giving people money helps them in the short term, but it doesn’t provide growth. It’s much more difficult to help people increase their value.We try to motivate people to take action and to improve their own results, their own lives. But that motivation lasts only a short time. It’s more difficult to help people identify and activate their real purpose, the internal motivation that they can sustain over their lifetime.We use quick fix solutions to long term, systemic problems. The person who needs help might take the help and go away, but they haven’t left any better than off then when you found them. They’re still dependent.Why You UnderinvestUnderinvesting produces dependents instead of independence. And there are five reasons that you might be underinvesting.Your failure to invest in people might mean that you believe the person you are “helping” can’t improve or grow. They aren’t worth investing in because you believe they are incapable of improving.People will surprise you if you give them a chance.It might also suggest that you believe they aren’t really worth the level of investment required, that you believe that there is something more important than helping them grow.Your legacy is going to consist of nothing more than the lives you touched, for good or for ill.It might be that you really don’t know how to give people the real help they need. This is probably the least likely reason; it’s likely that you are extremely capable and well suited to providing that level of help.You have the abilities you possess for a reason.Some people fail to give people the help they really need because it gives them power or control. If their help is always needed, then they are significant, they’re important.Nothing makes you more significant that helping people grow.But when it comes to helping people grow, the most likely reason you are underinvesting is that you aren’t really focused on that outcome. You’re focused on treating the presenting problem and not the real source of the problem. Changing the outcome and investing in caring enough to really help solves that problem—and it provides real help.What to InvestInvesting in people is expensive. Really helping requires you to make much greater investments.You have to make a greater investment of the most limited and fleeting of all your possessions: your time. Investing your time to really help produces outsized returns. The time invested to give people the help they need is returned to you many times over because it creates independence instead of dependents.You have to make a greater investment of your emotional energy. You have to care enough about the person that needs your help to give them the help that they really need. That means you have to shift the outcome from how you can help this person in the fastest way possible to how you can help them grow. It’s inefficient. That energy is expensive. It requires patience. It requires tolerance.If you are reading this and it has touched in you some way, then it’s likely there’s a long line of people that invested enough in caring about you to give you the help that you really needed. Remember what that looked like? Remember how that felt? Remember the difference it made in your life? That’s the investment in caring you need to make in helping others.QuestionsAre you giving the people that ask you for help the help that really need? Or are you doing what is efficient over what is really effective?Are you really creating the value that you are capable of creating? Or are you shortchanging yourself and the people that need your help?What is the root cause of your underinvestment in caring about other people enough to give them the help that they really need?last_img

Giving cities a road map to reducing their carbon footprint

first_imgCities are not just where 3.5 billion of us live—they are where more than  half of humanity uses electricity, drives cars, and throws out garbage, among myriad other activities that emit greenhouse gases. Now, a global coalition has released the first standardized method for measuring and reporting a given city’s greenhouse gas emissions. Called the Global Protocol for Community-Scale Greenhouse Gas Emission Inventories (GPC), the new standards were unveiled today at the United Nations’ ongoing climate negotiations in Lima.Cities are responsible for 70% of global carbon dioxide emissions, says Wee Kean Fong, who led development of the GPC at the World Resources Institute—a nonprofit based in Washington, D.C.—in partnership with the C40 Cities Climate Leadership Group and Local Governments for Sustainability (ICLEI). But there has been no standardized way to measure and report an individual city’s emissions. That has impeded plans to reduce urban climate footprints and track the effectiveness of local policies designed to reduce emissions. “You can’t manage what you can’t measure,” Fong says.A key element of the GPC is its recognition that a city may be responsible for gases emitted far outside its borders. Take power plants that burn fossil fuels to generate electricity, or landfills that receive solid waste, Fong says. Those can be located outside of a city, but their emissions are directly tied to urban activity. Holding cities accountable for such emissions may lead to some pushback when it comes to convincing them to adopt the GPC protocol, but it’s important for making sure measurements are accurate as possible, Fong says.Sign up for our daily newsletterGet more great content like this delivered right to you!Country *AfghanistanAland IslandsAlbaniaAlgeriaAndorraAngolaAnguillaAntarcticaAntigua and BarbudaArgentinaArmeniaArubaAustraliaAustriaAzerbaijanBahamasBahrainBangladeshBarbadosBelarusBelgiumBelizeBeninBermudaBhutanBolivia, Plurinational State ofBonaire, Sint Eustatius and SabaBosnia and HerzegovinaBotswanaBouvet IslandBrazilBritish Indian Ocean TerritoryBrunei DarussalamBulgariaBurkina FasoBurundiCambodiaCameroonCanadaCape VerdeCayman IslandsCentral African RepublicChadChileChinaChristmas IslandCocos (Keeling) IslandsColombiaComorosCongoCongo, The Democratic Republic of theCook IslandsCosta RicaCote D’IvoireCroatiaCubaCuraçaoCyprusCzech RepublicDenmarkDjiboutiDominicaDominican RepublicEcuadorEgyptEl SalvadorEquatorial GuineaEritreaEstoniaEthiopiaFalkland Islands (Malvinas)Faroe IslandsFijiFinlandFranceFrench GuianaFrench PolynesiaFrench Southern TerritoriesGabonGambiaGeorgiaGermanyGhanaGibraltarGreeceGreenlandGrenadaGuadeloupeGuatemalaGuernseyGuineaGuinea-BissauGuyanaHaitiHeard Island and Mcdonald IslandsHoly See (Vatican City State)HondurasHong KongHungaryIcelandIndiaIndonesiaIran, Islamic Republic ofIraqIrelandIsle of ManIsraelItalyJamaicaJapanJerseyJordanKazakhstanKenyaKiribatiKorea, Democratic People’s Republic ofKorea, Republic ofKuwaitKyrgyzstanLao People’s Democratic RepublicLatviaLebanonLesothoLiberiaLibyan Arab JamahiriyaLiechtensteinLithuaniaLuxembourgMacaoMacedonia, The Former Yugoslav Republic ofMadagascarMalawiMalaysiaMaldivesMaliMaltaMartiniqueMauritaniaMauritiusMayotteMexicoMoldova, Republic ofMonacoMongoliaMontenegroMontserratMoroccoMozambiqueMyanmarNamibiaNauruNepalNetherlandsNew CaledoniaNew ZealandNicaraguaNigerNigeriaNiueNorfolk IslandNorwayOmanPakistanPalestinianPanamaPapua New GuineaParaguayPeruPhilippinesPitcairnPolandPortugalQatarReunionRomaniaRussian FederationRWANDASaint Barthélemy Saint Helena, Ascension and Tristan da CunhaSaint Kitts and NevisSaint LuciaSaint Martin (French part)Saint Pierre and MiquelonSaint Vincent and the GrenadinesSamoaSan MarinoSao Tome and PrincipeSaudi ArabiaSenegalSerbiaSeychellesSierra LeoneSingaporeSint Maarten (Dutch part)SlovakiaSloveniaSolomon IslandsSomaliaSouth AfricaSouth Georgia and the South Sandwich IslandsSouth SudanSpainSri LankaSudanSurinameSvalbard and Jan MayenSwazilandSwedenSwitzerlandSyrian Arab RepublicTaiwanTajikistanTanzania, United Republic ofThailandTimor-LesteTogoTokelauTongaTrinidad and TobagoTunisiaTurkeyTurkmenistanTurks and Caicos IslandsTuvaluUgandaUkraineUnited Arab EmiratesUnited KingdomUnited StatesUruguayUzbekistanVanuatuVenezuela, Bolivarian Republic ofVietnamVirgin Islands, BritishWallis and FutunaWestern SaharaYemenZambiaZimbabweI also wish to receive emails from AAAS/Science and Science advertisers, including information on products, services and special offers which may include but are not limited to news, careers information & upcoming events.Required fields are included by an asterisk(*)A key selling point is that 35 cities have already benefited from implementing the GPC during its pilot phase in 2012. In the months since, other cities have been test-driving the new standards. David Maleki, a climate change analyst with the Inter-American Development Bank’s Emerging and Sustainable Cities Initiative, helps cities in Latin America and the Caribbean use the protocol to figure out which sectors are responsible for most of their greenhouse gas emissions. These “[greenhouse gas] inventories are a very basic building block for taking climate action in cities,” he says. Rio de Janeiro, for example, used a draft of the GPC to determine that transportation was responsible for a whopping 39% of the city’s total emissions; that led the government to focus on expanding public transit to more efficiently shrink its carbon footprint. “Building a greenhouse gas emissions inventory enables city leaders to manage their emissions reduction efforts, allocate resources and develop comprehensive climate action plans,” said Rio de Janeiro mayor and C40 Chair Eduardo Paes said in a statement.One lesson Fong and his team learned during the pilot program is that not every city is starting from the same place when it comes to measuring greenhouse gas emissions. Some cities, especially in the developing world, simply don’t have access to the kind of data needed for a comprehensive inventory; Maleki says he often works in places where the only emissions data available are on a national scale.To try to make the GPC accessible to cities that may not have all the right data, Fong’s team designed two tiers of reporting. Both incorporate transportation within the city, stationary burning of fossil fuels, consumption of electricity, and emissions related to waste. The more advanced tier adds data about industrial processes, land use change, transportation that brings people into the city, and other indirect sources of emissions. “The GPC is a very inclusive protocol,” says Ana Marques, project coordinator of ICLEI’s Urban-LEDS project, which will help cities in developing countries apply the GPC. “It enables all cities to participate.”last_img

Kobe sees dominant performance by Gilas Cadets in Kuala Lumpur meet

first_imgUPLB exempted from SEA Games class suspension K-pop star Jung Joon-young convicted of gang rape, spycam crimes Photo from Fiba.comNineteen-year-old basketball sensation Kobe Paras said Gilas Pilipinas will be a dominant force in the coming Southeast Asian Games and doesn’t mind taking a reduced role in the team’s gold-medal bid.“You can see the lineup, this is the best I’ve been part of,” said Paras, who has played in the US NCAA Division I, during a sendoff for Team Philippines Thursday night. “My teammates, when I was in high school, these guys are the ones I looked up to.”ADVERTISEMENT Catriona Gray spends Thanksgiving by preparing meals for people with illnesses Gilas Pilipinas routs Iraq, stays unscathed in Fiba Asia Cup Brace for potentially devastating typhoon approaching PH – NDRRMC MOST READ SEA Games in Calabarzon safe, secure – Solcom chief Trending Articles PLAY LIST 00:50Trending Articles00:50Trending Articles00:50Trending Articles01:37Protesters burn down Iran consulate in Najaf01:47Panelo casts doubts on Robredo’s drug war ‘discoveries’01:29Police teams find crossbows, bows in HK university01:35Panelo suggests discounted SEA Games tickets for students02:49Robredo: True leaders perform well despite having ‘uninspiring’ boss02:42PH underwater hockey team aims to make waves in SEA Games LATEST STORIES Read Next Typhoon Kammuri accelerates, gains strength en route to PH Don’t miss out on the latest news and information. Paras, whose father Benjie also won for the country a SEA Games gold, added that he will be taking on whatever role the coach requires of him.“You know I’m not selfish,” he said. “I won’t walk in and be like I wanna be the main scorer or anything like that. If coach tells me to defend, I’ll defend the best player out there. If he tells me to rebound, I’m gonna do it as much as I can.”FEATURED STORIESSPORTSWATCH: Drones light up sky in final leg of SEA Games torch runSPORTSSEA Games: Philippines picks up 1st win in men’s water poloSPORTSMalditas save PH from shutout View comments LOOK: Venues for 2019 SEA Games WATCH: Streetboys show off slick dance moves in Vhong Navarro’s weddinglast_img

Leave a Reply

Your email address will not be published. Required fields are marked *