This is one in a series of profiles showcasing some of Harvard’s stellar graduates.When Damon Clark ’17, a member of the Navajo (Diné) Nation, arrived on campus from New Mexico, he wore short hair, a cowboy hat, and hiking boots. He was here to absorb the best of Western education, to be transformed by the Harvard experience.At graduation, the “transformed” Clark plans to wear a traditional turquoise necklace and his moccasins, with his long black hair tied into the traditional bun called “tsiiyéél” in Navajo. In his culture, Clark noted, hair is considered an extension of a person’s thoughts and should not be cut.Clark will leave Harvard in May with both a greater knowledge of Navajo history and culture and a renewed pride in his indigenous identity.“I grew up learning that the U.S. government’s policy on Native Americans was, at some point, ‘Kill the Indian, Save the Man,’” said Clark, a Cabot House resident, his hair this day hanging in a single long braid. “They’d send Native Americans to boarding schools to ‘civilize’ them.“I have gone in the opposite direction,” he said. “At Harvard, I became more connected with my culture and identity. I learned about my history and how to remain true to myself.”Clark is in a small minority at Harvard, with 2.6 percent of students admitted last year were Native Americans. There are no full-time Native American faculty members on campus, said Shelly Lowe, executive director of the Harvard University Native American Program.In 1665, Caleb Cheeshahteaumuck of the Wampanoag tribe became the first Native American to graduate from Harvard. The Harvard Charter of 1650 stipulated the College’s commitment to “the education of the English & Indian youth of this country,” and Native American graduates number over 1,000 in 365 years.Clark, a social studies concentrator, took advantage of every opportunity on campus to raise awareness about his culture and the nuances of being a contemporary indigenous person, an endeavor that could be challenging.“I was the only Native American in many classes [and] the first Native American person many of my classmates had ever met,” Clark said. “That was daunting, at times, because I had to become an expert in Native American history and culture. I had to learn how to be a leader and represent both the Navajo people and Native Americans.”As part of that quest, Clark wore his moccasins (Kélchí in Navajo) everywhere he went, as a symbol of his identity. He wore them around campus, in winter, and after being awarded a Presidential Public Service Fellowship, he wore them to the White House in summer 2015 when he worked for the Office of Intergovernmental Affairs & Public Engagement as a Native American liaison intern.On his Instagram account, Clark posts close-up pictures of his moccasins in all the places he has visited. “I call them my Crimson moccasins,” he said, smiling. “I want to leave the footprints of my culture everywhere I go.”Clark is a great-great-grandson of Hastiin Ch’ilhaajinii, also known as Chief Manuelito, a Navajo leader who fought against the U.S. military and signed a treaty in 1868 to establish a reservation for his people. His grandmother is renowned Navajo master weaver Irene Clark, and his parents are both educators. Clark descends from the Near the Water clan on his mother’s side, and from the Near the Water’s Edge clan on his father’s.Clark grew up in Asaayii, Bowl Canyon, a rural community in the home of Diné Nation. With a size similar to West Virginia, touching on New Mexico, Utah, and Arizona, it is the largest Indian reservation in the United States, and with more than 300,000 enrolled members, Navajos are the largest Indian tribe in the nation.Like many Navajos, Clark’s family lives off the grid without plumbing or electricity. Clark has gone home every summer to visit his parents, to “be a farmer” and be connected to his roots.“When my father became a doctor of philosophy, my grandfather told him not to get too smart for his own good,” said Clark. “He said, ‘A Navajo man can be smart, but he has to know how to farm, chop wood, and build a fire.’”Last winter, Clark led a 10-day trip for eight undergrads to experience the Navajo way of life and perform public service in his community. Students chopped wood for elders and needy families, met with high school students and tribal leaders, and attended classes in Navajo. The trip was a success, said Clark, and the students that went became his “brothers and sisters.”Clark also led a powwow on Harvard’s campus and requested that the College offer a Navajo language course so that he could work on his thesis on Navajo entrepreneurship. The added course was the first of its kind offered on campus, and Clark cherishes it as one of the highlights of his time at Harvard.Another highlight, a testament to Clark’s will and intelligence, is having learned how to swim in a week in order to make the crew team. “There aren’t pools where I’m from,” he said. “But I wanted to learn, and I googled it.”After graduation, Clark will travel to New Zealand, where he will spend a year studying the Maori indigenous people. “I’m looking forward to leaving my footprints in New Zealand,” he said.Like many Native Americans in the United States, Clark straddles two worlds, that of his ancient culture and that of his country of birth. He hopes to serve as a bridge between them, meanwhile following the advice of his great-great-grandfather, Chief Manuelito.“He said to the Navajo people, ‘Go, my children, climb the ladder of education,’” said Clark. “But he also talked about remaining true to our culture and our identity. That is one of my aspirations, to bring together Western education and Navajo culture for the benefit and empowerment of my community.”
Lately, I have spent large swaths of my time focused on Deep Learning and Neural Networks (either with customers or in our lab). One of the most common questions that I get is around underperforming model training with regard to “wall clock time.” This has more to do with focusing on only one aspect of their architecture, say GPUs. As such, I will spend a little time writing about the 3 fundamental tenets for a successful Deep Learning architecture. These fundamental tenants are: compute, file access, and bandwidth. Hopefully, this will resonate and help provide some thoughts for those customers on their journey.OverviewDeep Learning (DL) is certainly all the rage. We are defining DL as a type of Machine Learning (ML) built on a deep hierarchy of layers, with each layer solving different pieces of a complex problem. These layers are interconnected into a “neural network.”The use cases that I am presented with continue to grow exponentially with very compelling financial return on investments. Whether it is Convolutional Neural Networks (CNNs) for Computer Vision or Recurrent Neural Networks (RNNs) for Natural Language Processing (NLP) or Deep Belief Networks (DBN) for Restricted Boltzmann Machines (RBMs), Deep Learning has many architectural structures and acronyms. There is some great Neural Network information out there. Pic 1 is a good representation of the structural layers for Deep Learning on Neural Networks:OrchestrationOrchestration tools like BlueData, Kubernetes, Mesosphere, or Spark Cluster Manager are the top of the layer cake of implementing Deep Learning with Neural Networks. These provide scheduler and possibility container capabilities to the stack. This layer is the most visible to the Operations team running the Deep Learning environment. There are certainly pros and cons to the different orchestration layers, but that is a topic for another blog.Deep Learning FrameworksCaffe2, CNTK, Pikachu, PyTorch, or Torch. One of these is a cartoon game character. The rest sound like they could be in a game, but they are some of the blossoming frameworks that support Deep Learning with Neural Networks. Each framework has their pros and cons with different training libraries and different neural networks structures (s) for different use cases. I regularly see a mix of frameworks within Deep Learning environments and the Framework chosen rarely changes the 3 tenets for architecture.Architectural TenetsI’ll use an illustrative use case to highlight the roles of the architectural tenets below. Since the Automotive industry has Advanced Driving (ADAS) and Financial Services have Trader Surveillance use cases, we will explore a CNN with Computer Vision. Assume a 16K resolution image that stores around 1 gigabyte (GB) in a file on storage and has 132.7 million pixels.ComputeTo dig right in, the first architectural tenet is Compute. The need for compute is one of those self-obvious elements of Deep Learning. Whether you use GPU, CPU, or a mix tends to result from which neural network structure (CNNs vs RNNs vs DBNs), use cases, or preferences. The internet is littered with benchmarks postulating CPUs vs GPUs for different structures and models. GPUs are the mainstay that I regularly see for Deep Learning on Neural Networks, but each organization has their own preferences based on past experiences, budget, data center space, and network layout. The overwhelming DL need for Compute is for lots of it.If we examine our use case of the 16K image, the CNN will dictate how the image is addressed. The Convolutional Layer or the first layer of a CNN will parse out the pixels for analysis. 132.7M pixels will be fed to 132.7M different threads for processing. Each compute thread will create an activation map or feature map that helps to weight the remaining CNN layers. Since this volume of threads for a single job is rather large, the architecture discussion around concurrency versus recursion of the neural network certainly evolves from the compute available to train the models.Embarrassingly ParallelIf we start with the use case, this paints a great story to start with for file access. We already discussed that a 16K resolution image will spawn 132.7 million threads. What we didn’t discuss is that these 132.7 million threads will attempt to read the same 1 GB file. Whether that is at the same time or over a time window depends on the amount of compute available for the model to train with. In a large enough compute cluster, those reads can be simultaneous. This scenario is referred to being “embarrassingly parallel” and there are great sources of information on it. Pic 2 denotes the difference between regular command and control on high performance computing workloads (HPC) with “near embarrassingly parallel” vs embarrassingly parallel in Deep Learning.In most scale up storage technologies, embarrassingly parallel file requests lead to increased latency as more threads open the file. This eventually leads to a logarithmic asymptote that approaches infinity with enough file opens. This means that the more threads that open will often never complete until the concurrency level on the file open is reduced.In true scale out technologies, embarrassingly parallel file opens are a mathematical function of bandwidth per storage chassis and number of opens requested per neural network structure.Massive BandwidthI am often told that latency matters in storage. I agree for certain use cases. I do not agree with Deep Learning. Latency is a single stream function of a single process. When 132.7M files read the same file in an embarrassingly parallel fashion, it is all about the bandwidth. A lack of significant forethought into how the compute layer gets “fed” with data is the biggest mistake I see in most Deep Learning architectures. It accounts for most of the wall clock time delays that customers focus on.While there is no right answer as to what constitutes “fast enough” for feeding the Deep Learning structures, there certainly is “good enough”. Good enough usually starts with a scale out storage architecture that allows a great mix of spindle to network feeds. 15GB per second for a 4 rack unit chassis with 60 drives is a good start.Wrap UpIn summary, Deep Learning with Neural Networks is a blossoming option in the analytics arsenal. Its use cases are growing quite regularly with good results. The architecture should be approached holistically though instead of just focusing on one aspect of the equation. The production performance of the architecture will suffer depending on which tenet is skimped on. We regularly have conversations with customers around their architectures and welcome a more in-depth conversation around your journey. If you would like more details on how Dell EMC can help you with Deep Learning, feel free to email us at firstname.lastname@example.org
A growing market The Sinaloa Cartel has forged alliances with two Hong Kong-based mafias to acquire precursor chemicals which are used to manufacture highly-addictive synthetic drugs, authorities said. The Sinaloa Cartel is obtaining precursor chemicals from 14K and Sun Yee On, two Chinese triads – organized crime groups — which are based in Hong Kong, the South China Morning Post reported recently. The cartel, which is led by fugitive kingpin Joaquin “El Chapo” Guzman, uses the chemicals to manufacture methamphetamines in Guatemala and El Salvador, officials said. Sinaloa Cartel operatives, including local gangs, pick up the precursor chemicals from ports in Guatemala and Honduras, then drive the chemicals in SUVs to clandestine labs. Once the synthetic drugs are processed, Sinaloa Cartel operatives transport them to illegal drug markets in the United States, Canada, Europe, Africa and Mexico. Through alliances with local gangs, the cartel also sells some of the synthetic drugs in Argentina, Guatemala and El Salvador. Ties to ‘El Chapo’ Law enforcement agents in the Philippines and in China recently discovered evidence confirming the alliance between El Chapo and the Hong Kong mafias. On Dec. 27, 2013, law enforcement agents from the Philippines, working in cooperation with U.S. security forces, captured three suspects near Manila. The three suspects have ties to the Sinaloa Cartel, authorities said. The three suspects were preparing and storing methamphetamine in a laboratory located on a rooster farm south of Manila, authorities said. Security forces seized 84 kilos of methamphetamine. The precursor chemicals used to produce the synthetic drugs were provided by 14K and Sun Yee On, The three suspects had ties to the Sinaloa Cartel, authorities said. Bartolome Tobia, the chief of Filipino anti-narcotics police forces, said the security operation confirmed that El Chapo was working with Chinese mafias. “Mexicans, who are already here, are getting help from Chinese criminal organizations linked to drug trafficking,” he said. A week after the arrests in the Philippines, on Jan. 4, 2014, more than 2,500 Chinese security agents seized more than three tons of methamphetamine in the city of Guangdong Lugeng. Chinese authorities also arrested 182 suspects. The synthetic drugs belonged to one of the Chinese organized crime groups that is working with El Chapo, authorities said. Organized crime is a transnational and global phenomenon, and borders are non-existent for the Sinaloa Cartel and the Chinese triads, Chabat said. “The drug cartels are fighting to control a single territory, the world,” Chabat said. In particular, El Chapo is trying to control drug trafficking not only in Mexico, but in the U.S., Europe, all of Latin America, Africa, Asia and Australia, aas part of a “deviant globalization strategy,” according to insightcrime.org. “There is no doubt that the Sinaloa Cartel is one of the strongest and most powerful drug trafficking organizations on the American continent,” Chabat said. Authorities in Asian countries have captured several of El Chapo’s operatives in recent years. For example, in June 2011, Malaysian authorities sentenced three Mexicans from Sinaloa to death for operating a methamphetamine laboratory in the city of Johor. Malaysian police arrested the brothers Luis Alfonso, Simón and José Regino González Villareal, alleged members of the Sinaloa Cartel, in 2008. . In November 213, Latin American security officials agreed at the “Second Regional Conference on Prevention, Control and Police Investigation on the diversion of Precursor Chemicals,” to increase international cooperation in the battle against synthetic drugs. Latin American authorities pledged to strengthen corporate social responsibility and collaboration between the business sector and the state to enhance the control of precursor chemicals. They also agreed to improve various international programs and projects, as well as strengthen work on scientific evidence in order to convert information into knowledge for better tools in the fight against drugs. “As part of international cooperation, institutions should be strengthened in all countries at the international level in the fight against synthetic drugs,” Chabat said. Regarding the return of the Colombian ceramics, I wonder if they returned everything they said there was o like during the colonial period, they kept the gold and the silver and they returned what was worthless. Because the Europeans greed for what is ours is well known. Borders not relevant to ‘El Chapo’ and Chinese triads In recent years, the market for amphetamine-type synthetic drugs has been expanding. Authorities in the U.S. and other countries where the drugs are sold are reporting record levels of synthetic drug seizures and use. About 80 percent of the methamphetamine consumed in the U.S. each year is prepared with precursor chemicals from China, authorities estimate. Argentina, El Salvador, Guatemala, Mexico, and United States are among the countries with highest number of drug processing laboratories, authorities have estimated. Synthetic drugs can be swallowed, injected, smoked or inhaled and are also known by popular names such as “crystal”, “glass”, “cocaine of the poor”, “yaba”, “ice” or “meth.” These drugs are highly addictive and can cause mental disorders such as schizophrenia, paranoia and aggressive behavior. The governments of Mexico and the United States strictly control the sale of precursor chemicals. These controls have prompted the Sinaloa Cartel and other Latin American drug traffickers to forge alliances with Asian organized crime groups, Chabat said. Relatively weak regulation in the Chinese chemical industry provides local and international drug traffickers with easy access to psychoactive substances in large quantities, he added. Hong Kong has not issued certificates guaranteeing that the people who purchase chemical precursors are the actual recipients of the precursors, according to the United Nations. A history of selling synthetic drugs By Dialogo February 19, 2014 The Sinaloa Cartel has sold methamphetamine since the 1990s. In the early to mid-1990s, Ignacio “El Nacho” Coronel, who at the time was one of El Chapo’s top lieutenants, “clearly saw” the great potential in the emerging market for methamphetamine, according to the “Security and Defense Atlas of Mexico 2012,” which was published by the Mexico City-based Collective for the Analysis of Security with Democracy (CASEDE). El Nacho, who also came to be known as “The King of Crystal,” was in charge of producing and trafficking methamphetamine for El Chapo. He developed a system to transport large amounts of ephedrine, which is used to make methamphetamine, from Asia to Mexico, where cartel operatives processed synthetic drugs. Mexican Army soldiers killed El Nacho during a gun battle in Zapopan, Jalisco, in July 2010. The alliances between the Sinaloa Cartel and the two Chinese organized crime groups are “deadly,” said Jorge Chabat, a security analyst at the Center for Research and Teaching in Economics (CIDE) in Mexico City. The Sinaloa Cartel is the largest, most violent transnational criminal organization in the world, Chabat explained. The two Chinese mafias are major producers of precursor chemicals and methamphetamines, Chabat said. The new alliances will help the forces of El Chapo produce and transport a steady supply of synthetic drugs, Chabat said. “The Sinaloa Cartel will obtain precursors in large amounts, reduce costs, and continue to internationalize,” Chabat said. “The U.S. market for synthetic drugs is attractive to ‘El Chapo’ and the Chinese triads.” Triads are Chinese criminal organizations based in Hong Kong, Taiwan, and mainland China. The 14K triad, has about 25,000 members in its ranks. The other traid El Chapo is allied with, Sun Yee On, has more than 50,000 members. El Chapo is using substances such as phenyl ethyl acetate and isobutyl phenylacetate to try to create a new synthetic drug, according to published reports. Deadly alliances
As a child, I would sit for hours on the floor with a freshly emptied box of Legos. Sometimes, I would start by sorting out all the pieces and flipping through the instruction manual to gain an understanding of the plans and steps I was going to take.Other times, I would take more of a free flow approach and just start building whatever I wanted. When this happened, I would often feel overwhelmed and become frustrated when the end product did not turn out as I had envisioned it in my mind’s eye.Now as a father, I am guiding my kids to build structures greater than I ever did. And like me, they experience higher levels of success and feelings of gratitude when they have a guide to help them along the way.The same is true for financial digital marketing. You too are giving two paths. You can take the first path and have a blueprint to help guide you along the way. Or alternatively, you can take the free flow approach. Up to this point, many financial institutions have chosen the second path and just start building without having an understanding of where they are headed. continue reading » 2SHARESShareShareSharePrintMailGooglePinterestDiggRedditStumbleuponDeliciousBufferTumblr
continue reading » As you age, you tend to think more about your eyesight and vision. It’s only natural that the older we get the quality of how we see tends to diminish. Whether you are already wearing corrective lenses or “readers” it’s only a matter of time before your lenses get cloudy. In other words, it’s hard to maintain 20/20 vision your entire life.Maintaining a clear 20/20 vision for your credit union or bank is no different. Over time it’s extremely easy for your strategic vision to also get a little cloudy. Things happen: internal staffing challenges arise, external competitors pop up and unexpected roadblocks all cause us to take our eyes off our vision.But as we enter the new year, now is a perfect time to reflect on how to maintain a clear 20/20 strategic vision.Here are three tips for making sure your 2020 vision is 20/20: ShareShareSharePrintMailGooglePinterestDiggRedditStumbleuponDeliciousBufferTumblr
To access this article REGISTER NOWWould you like print copies, app and digital replica access too? SUBSCRIBE for as little as £5 per week. Would you like to read more?Register for free to finish this article.Sign up now for the following benefits:Four FREE articles of your choice per monthBreaking news, comment and analysis from industry experts as it happensChoose from our portfolio of email newsletters
The State Palace was not immediately available for comment when asked about Jokowi’s call with Rouhani. But the Iranian Embassy in Jakarta has confirmed it.On Tuesday, Jokowi stressed on his official Twitter account the need for solidarity among nations during the COVID-19 pandemic.“COVID-19 knows no borders, ethnicities and religions. It is impossible to fight it alone. That’s why I have been talking to several leaders of countries such as Saudi Arabia, the UAE, the United States, China, Russia, Japan, Singapore, Australia and others, in the spirit of solidarity and helping each other,” he said. The leaders also talked about the COVID-19 pandemic and the two countries’ plans to fight the highly contagious disease. During the telephone conversation, Jokowi and Rouhani also shared each country’s experiences and scientific achievements as well as plans to develop cooperation to meet each other’s needs.”The Islamic Republic of Iran has achieved good success in the field of science and technology and supplying and manufacturing requirements such as test kits, ventilators, CT scans and N95 masks and is ready to cooperate and interact with Indonesia in this regard,” Rouhani said as quoted by ISNA on Monday.As of Wednesday, Iran has recorded 92,584 cases of COVID-19, with 5,877 deaths, Worldometers recorded.Meanwhile, Indonesia has reported 9,511 confirmed cases with 773 fatalities. Topics : President Joko “Jokowi” Widodo had a fruitful conversation over the phone with Iranian President Hassan Rouhani on Monday, during which they stressed the importance of cooperation in the fight against COVID-19.The two leaders acknowledged the friendly relationship between their countries and highlighted the further development of ties and cooperation between Tehran and Jakarta to serve both nations’ interests, Iranian news outlet ISNA reported.Rouhani also congratulated Indonesians on the holy month of Ramadan, expressing hope that it would bring only blessings for all Muslims.
Half of the large listed companies in Germany still have much to do when it comes to the governance of their pension funds, according to Susanne Jungblut, director of Solution Compensation & Benefits at KPMG Germany.Jungblut also warned that, judging by the results of a recent survey of one-third of those companies listed in the DAX and MDAX indices with more than €100m in pension liabilities the “need for action” was likely to be much higher among smaller companies, which “still often have large defined benefit obligations” on their balance sheets.“Especially in the low-interest-rate environment, risk management of pension liabilities has significant importance and should be tackled,” she said at the Handelsblatt conference in Berlin.The survey found that, while 90% of the companies surveyed regularly run ALM studies, only around half of those do so quarterly or monthly. Similarly, the legally required ‘sensitivity analysis’ of liabilities is conducted only once a year by one-third of the companies, while half of respondents conceded they did not know how often such analyses were conducted.“This is not sufficient given the fast economic developments,” Jungblut said.She added that, while most companies actively include their actuaries in assessing pension risks, she was worried about the minority where the actuary seemed to be the only one with knowledge of that risk.Jungblut also said it was “really bad” that a minority of companies lacked any defined responsibilities regarding pension risks.Special reporting tools for pensions, a pensions committee and a pensions governance directive were each found in half of the surveyed companies, respectively.Jungblut noted that German companies had “become aware of the relevance of pension governance and had also recognised the need for action”.
Adwen has hired Eidesvik’s vessel Viking Neptun to support commissioning activities at the 350MW Wikinger offshore wind farm in the German Baltic Sea.The vessel will be used for the accommodation of technicians and crew transfers to the wind turbines.The contract, beginning from mid-October, has the duration of five months with further options.Jan Fredrik Meling, CEO at Eidesvik, said: “I am very pleased with our organisation’s ability to find employment for the Viking Neptun in the renewable market. This is the fourth contract Eidesvik has been awarded in the offshore wind market since the beginning of 2016.”At the end of August, 56 out of 70 5MW wind turbines were installed at the Wikinger wind farm, while all 70 are scheduled to be up and connected to the grid before the year end.The EUR 1.4 billion wind farm will be operated from the recently opened O&M base in Mukran for at least 25 years.Wikinger is scheduled to be fully operational by the end of this year.
Sweden-based shipowner Furetank Rederi, a part of the Gothia Tanker Alliance, has taken delivery of its new chemical/product tanker, the 20,000 cbm Fure Vinga.Featuring a length of 149.9 meters and a beam of 22.8 meters, the newbuilding was delivered by China’s shipbuilder Avic Dingheng Shipbuilding.The tanker will be commercially managed by Furetank Chartering in the Gothia Tanker Alliance.The 16,300 dwt tanker is a part of a batch of six vessels ordered by Furetank Rederi together with the alliance partners Thun Tankers and Älvtank Rederi.Being built to a design developed by FKAB together with Furetank, the vessels have a minimal impact on the environment, with close to 50% reduction of CO2 emissions.They fulfill the Tier III rules and have LNG for inert gas production, power production with floating frequency, battery backup (UPS) for all vital functions to minimize use of auxiliary engines, installed ballast water cleaning system, ice class 1A and Alternative Propulsion System.