02 outubro 2016
01 outubro 2016
Deep Learning: passado, presente e futuro
Excelente reportagem da Fortune sobre o Deep Learning.
But what most people don’t realize is that all these breakthroughs are, in essence, the same breakthrough. They’ve all been made possible by a family of artificial intelligence (AI) techniques popularly known as deep learning, though most scientists still prefer to call them by their original academic designation: deep neural networks.
The most remarkable thing about neural nets is that no human being has programmed a computer to perform any of the stunts described above. In fact, no human could. Programmers have, rather, fed the computer a learning algorithm, exposed it to terabytes of data—hundreds of thousands of images or years’ worth of speech samples—to train it, and have then allowed the computer to figure out for itself how to recognize the desired objects, words, or sentences.
In short, such computers can now teach themselves. “You essentially have software writing software,” says Jen-Hsun Huang, CEO of graphics processing leader Nvidia NVDA 1.66% , which began placing a massive bet on deep learning about five years ago. (For more, read Fortune’s interview with Nvidia CEO Jen-Hsun Huang.)
Neural nets aren’t new. The concept dates back to the 1950s, and many of the key algorithmic breakthroughs occurred in the 1980s and 1990s. What’s changed is that today computer scientists have finally harnessed both the vast computational power and the enormous storehouses of data—images, video, audio, and text files strewn across the Internet—that, it turns out, are essential to making neural nets work well. “This is deep learning’s Cambrian explosion,” says Frank Chen, a partner at the Andreessen Horowitz venture capital firm, alluding to the geological era when most higher animal species suddenly burst onto the scene.
That dramatic progress has sparked a burst of activity. Equity funding of AI-focused startups reached an all-time high last quarter of more than $1 billion, according to theCB Insights research firm. There were 121 funding rounds for such startups in the second quarter of 2016, compared with 21 in the equivalent quarter of 2011, that group says. More than $7.5 billion in total investments have been made during that stretch—with more than $6 billion of that coming since 2014. (In late September, five corporate AI leaders—Amazon, Facebook, Google, IBM, and Microsoft—formed the nonprofitPartnership on AI to advance public understanding of the subject and conduct research on ethics and best practices.)
Google had two deep-learning projects underway in 2012. Today it is pursuing more than 1,000, according to a spokesperson, in all its major product sectors, including search, Android, Gmail, translation, maps, YouTube, and self-driving cars. IBM’s IBM 0.47% Watson system used AI, but not deep learning, when it beat two Jeopardy champions in 2011. Now, though, almost all of Watson’s 30 component services have been augmented by deep learning, according to Watson CTO Rob High.
Venture capitalists, who didn’t even know what deep learning was five years ago, today are wary of startups that don’t have it. “We’re now living in an age,” Chen observes, “where it’s going to be mandatory for people building sophisticated software applications.” People will soon demand, he says, “ ‘Where’s your natural-language processing version?’ ‘How do I talk to your app? Because I don’t want to have to click through menus.’ ”
[...]
Fato da Semana: Contas do setor público
Data: 30 setembro de 2016
Descrição do Fato
O déficit primário de agosto foi de 22 bilhões. Em doze meses o déficit é de 169 bilhões, fortemente influenciado pelo resultado de dezembro.
Explicações para o Déficit
Existem diversas causas para o déficit do setor público. Podemos citar entre elas:
A redução da receita do governo, em razão da recessão e da concessão de uma série de benefícios para empresas e setores, da frustração na venda de ativos e da redução dos dividendos das empresas estatais.
Durante os últimos anos os salários do pessoal do setor público cresceu bem acima da inflação e do setor privado. Além disto, aumentaram o número de pessoas contratadas. Na realidade, o setor público não consegue obter ganhos de produtividade.
A sociedade brasileira fez uma opção de participação do estado em setores cujo crescimento dos gastos é tradicionalmente superior aos outros setores. Vide os ensinamentos de Baumol sobre este assunto.
As empresas estatais são ineficientes e a participação do setor público na economia é desastroso, incentivando ineficiências e gastos questionáveis. (Isto inclui a nossa previdência)
Notícia boa para contabilidade? Não
A contabilidade pública parece não ajudar a sociedade a controlar o tamanho do estado, impedindo crises como esta. Pelo contrário, a nossa contabilidade pública é obscura, atrasada, legalista, pouco informativa, baseada numa teoria ultrapassada e com pouca perspectiva de mudança deste perfil.
Desdobramentos
Com respeito ao déficit, levaremos anos para corrigir os erros grosseiros cometidos nos últimos anos. Não consigo ver algo positivo sendo apresentado nos próximos anos.
Mas a semana só teve isto?
A crise do banco alemão Deustche, que não começou agora, mas está se agravando, talvez traga consequências serias para a economia mundial.
30 setembro 2016
Previsão do mercado
O mercado tem sido regularmente usado como uma forma de previsão.No Brasil nós temos os casos do Focus, do Banco Central, que tenta simular um "mercado de especialistas" para projeção de variáveis econômicas. Ou o Prisma Fiscal, mais recentemente, para variáveis das finanças públicas. Pela sabedoria das massas, o mercado seria um bom previsor.
O gráfico abaixo é da Iowa Eletronic Market para as eleições dos EUA. O mercado de apostas indica que Clinton tem 72% de chance de vencer as eleições.
O gráfico abaixo é da Iowa Eletronic Market para as eleições dos EUA. O mercado de apostas indica que Clinton tem 72% de chance de vencer as eleições.
Oi e a recuperação judicial
Segundo informações do Financista, a empresa de auditoria PwC e o escritório de advocacia Wald estão pedindo mais de 300 milhões de reais para gerenciar a recuperação judicial da empresa de telefonia Oi. A proposta foi entregue na Vara responsável pelo caso. Do valor total, 214 milhões seriam a parte da PwC.
29 setembro 2016
Rei emérito da Espanha e a tentativa de suborno
O rei emérito da Espanha, Juan Carlos I, provavelmente tentou subornar um grupo anticorrupção. Diante da possibilidade da princesa Cristina ser a primeira pessoa da família real espanhola a ser julgada criminalmente, o rei supostamente ofereceu dinheiro, dois milhões de euros, para que o assunto fosse esquecido. Um gravação mostrou uma conversa entre um intermediário do rei que solicitava a organização Manos Limpias retirar a acusação de fraude fiscal contra a princesa, revelou o jornal El Mundo (via aqui).
Além da princesa, o ex-atleta Urdangarin, marido de Cristina, também está sendo acusado. O juiz ainda não deu veredito do caso.
Em 2014 o rei abdicou do trono e em seu lugar assumiu o rei Felipe. A conversa gravada é do final do ano passado.
Além da princesa, o ex-atleta Urdangarin, marido de Cristina, também está sendo acusado. O juiz ainda não deu veredito do caso.
Em 2014 o rei abdicou do trono e em seu lugar assumiu o rei Felipe. A conversa gravada é do final do ano passado.
Neymar
Neymar entrou com uma ação contra a União Federal pelo vazamento do processo no qual ele e seus pais são investigados por suposta sonegação de impostos no Brasil. O jogador reclama que o acesso ao processo é sigiloso, porém informações foram divulgadas na imprensa sem a sua autorização. (Fonte: aqui)
A acusação é sonegação de impostos e ocorreu entre 2011 a 2013.
A acusação é sonegação de impostos e ocorreu entre 2011 a 2013.
28 setembro 2016
Links
Value Walk: Goodbye Dilma, Hello Michel
Em 1906 antecipou os problemas da tecnologia moderna
A importância do emprego seguro para a pesquisa universitária
Como os lobos mudam os rios
IFRS para pequenas empresas
Cidades sustentáveis (Zurique em primeiro, SP em 84o. e RJ em 86o.)
Economizando no café
Fonte: Aqui
Já conhecia um trabalho próximo, na UnB, sobre o consumo do café, realizado pelo Tiago Mota. Agora na UFRN. Pequenos gastos, grandes economias.
O futuro do papel
Essa tendência indica que avanços estão sendo feitos em várias frentes, desde empresas de tecnologia como a DocuSign Inc., a maior participante do mercado de assinaturas eletrônicas, à ascensão de tablets e de dispositivos móveis. Mais importante, ela representa uma mudança que levou muito mais tempo do que o previsto. Ela foi atrasada pelo fato de que os negócios são realizados de formas muito mais complicadas do que qualquer um gostaria. A persistência do papel no local de trabalho — 60% do qual não é impressão opcional, diz Shane — significa que os processos empresariais mudam devagar, se é que mudam. As pequenas e médias empresas são as mais lentas em se livrar do papel, ou seja, para digitalizar totalmente seu fluxo de trabalho.Também existe o fato de que papel é fabuloso. É a única tecnologia de exibição e inserção de dados que custa pouco, não pesa quase nada, é legível com quase qualquer tipo de luz e não exige uma conexão de internet. É a personificação da portabilidade e da durabilidade. (WSJ. Continue lendo aqui)
Na UnB adotamos este ano o processo eletrônico. Não se tem um levantamento do impacto, mas é perceptível a economia de papel. (Foto: Aqui)
27 setembro 2016
Desenho de mecanismos: os economistas como engenheiros
For eight years, Peter Coles had an economist’s dream job at Harvard Business School.
His research focused on the design of efficient markets, an important and growing field that has influenced such things as Treasury bill auctions and decisions on who receives organ transplants. He even got to work with Alvin E. Roth, who won a Nobel in economic science in 2012.
But prestige was not enough to keep Mr. Coles at Harvard. In 2013, he moved to the San Francisco Bay Area. He now works at Airbnb, the online lodging marketplace, one of a number of tech companies luring economists with the promise of big sets of data and big salaries.
Silicon Valley is turning to the dismal science in its never-ending quest to squeeze more money out of old markets and build new ones. In turn, the economists say they are eager to explore the digital world for fresh insights into timeless economic questions of pricing, incentives and behavior.
“It’s an absolute candy store for economists,” Mr. Coles said.
The pay, of course, is a lot better than you would find in academia, where economists typically earn $125,000 to $150,000 a year. In tech companies, pay for a Ph.D. economist will usually come in at more than $200,000 a year, the companies say. With bonuses and stock grants, compensation can easily double in a few years. Senior economists who manage teams can make even more.
Businesses have been hiring economists for years. Usually, they are asked to study macroeconomic trends — topics like recessions and currency exchange rates — and help their employers deal with them.
But what the tech economists are doing is different: Instead of thinking about national or global trends, they are studying the data trails of consumer behavior to help digital companies make smart decisions that strengthen their online marketplaces in areas like advertising, movies, music, travel and lodging.
Tech outfits including giants like Amazon, Facebook, Google and Microsoftand up-and-comers like Airbnb and Uber hope that sort of improved efficiency means more profit.Continue reading the main story
At Netflix, Randall Lewis, an economic research scientist, is finely measuring the effectiveness of advertising. His work also gets at the correlation-or-causation conundrum in economic behavior: What consumer actions occur coincidentally after people see ads, and what actions are most likely caused by the ads?
At Airbnb, Mr. Coles is researching the company’s marketplace of hosts and guests for insights, both to help build the business and to understand behavior. One study focuses on procrastination — a subject of great interest to behavioral economists — by looking at bookings. Are they last-minute? Made weeks or months in advance? Do booking habits change by age, gender or country of origin?
“They are microeconomic experts, heavy on data and computing tools like machine learning and writing algorithms,” said Tom Beers, executive director of the National Association for Business Economics.
Understanding how digital markets work is getting a lot of attention now, said Hal Varian, Google’s chief economist. But, he said, “I thought it was fascinating years ago.”
Mr. Varian, 69, is the godfather of the tech industry’s in-house economists. Once a well-known professor at the University of California, Berkeley, Mr. Varian showed up at Google in 2002, part time at first, but soon became an employee. He helped refine Google’s AdWords marketplace, where advertisers bid to have their ads shown on search pages, based on the words users type into Google’s search engine.
But prestige was not enough to keep Mr. Coles at Harvard. In 2013, he moved to the San Francisco Bay Area. He now works at Airbnb, the online lodging marketplace, one of a number of tech companies luring economists with the promise of big sets of data and big salaries.
Silicon Valley is turning to the dismal science in its never-ending quest to squeeze more money out of old markets and build new ones. In turn, the economists say they are eager to explore the digital world for fresh insights into timeless economic questions of pricing, incentives and behavior.
“It’s an absolute candy store for economists,” Mr. Coles said.
The pay, of course, is a lot better than you would find in academia, where economists typically earn $125,000 to $150,000 a year. In tech companies, pay for a Ph.D. economist will usually come in at more than $200,000 a year, the companies say. With bonuses and stock grants, compensation can easily double in a few years. Senior economists who manage teams can make even more.
Businesses have been hiring economists for years. Usually, they are asked to study macroeconomic trends — topics like recessions and currency exchange rates — and help their employers deal with them.
But what the tech economists are doing is different: Instead of thinking about national or global trends, they are studying the data trails of consumer behavior to help digital companies make smart decisions that strengthen their online marketplaces in areas like advertising, movies, music, travel and lodging.
Tech outfits including giants like Amazon, Facebook, Google and Microsoftand up-and-comers like Airbnb and Uber hope that sort of improved efficiency means more profit.Continue reading the main story
At Netflix, Randall Lewis, an economic research scientist, is finely measuring the effectiveness of advertising. His work also gets at the correlation-or-causation conundrum in economic behavior: What consumer actions occur coincidentally after people see ads, and what actions are most likely caused by the ads?
At Airbnb, Mr. Coles is researching the company’s marketplace of hosts and guests for insights, both to help build the business and to understand behavior. One study focuses on procrastination — a subject of great interest to behavioral economists — by looking at bookings. Are they last-minute? Made weeks or months in advance? Do booking habits change by age, gender or country of origin?
“They are microeconomic experts, heavy on data and computing tools like machine learning and writing algorithms,” said Tom Beers, executive director of the National Association for Business Economics.
Understanding how digital markets work is getting a lot of attention now, said Hal Varian, Google’s chief economist. But, he said, “I thought it was fascinating years ago.”
Mr. Varian, 69, is the godfather of the tech industry’s in-house economists. Once a well-known professor at the University of California, Berkeley, Mr. Varian showed up at Google in 2002, part time at first, but soon became an employee. He helped refine Google’s AdWords marketplace, where advertisers bid to have their ads shown on search pages, based on the words users type into Google’s search engine.
[...]
Fonte: aqui
26 setembro 2016
Assinar:
Comentários (Atom)
Postagens
