Quoc Le and His Quest for the Ultimate AI
Apr06

Quoc Le and His Quest for the Ultimate AI

Machines are very good at a great many things. Indeed, computers are better with numbers than human beings could ever hope to be. They can process and store vast quantities of data, but they can only use it to complete operations for which they’ve been programmed to perform. They can’t actually learn from the information they’re given. Quoc Le aims to change that. Machines don’t interpret the world in the same ways that we do. Their input consists entirely of numbers, and they simply aren’t capable of the sort of abstract thought that our brains use to learn about the world. We can take a symbolic representation of a cat – like a drawing or a description – and use that information to identify a real cat – even if we’ve never seen one before. Before Le came along, machines couldn’t do this. He worked as one of the main coders behind the Google Brain, a system that was able to teach itself to recognize images of cats on YouTube. It’s the search giant’s venture into the realm of deep learning – a field of artificial intelligence aimed at creating machines that work in ways that mimic the human brain – and it’s just the beginning of work that could revolutionize computing. In addition to his work on Google Brain, Le also developed a system that maps words to vectors, turning them into unique sets of numbers from which a computer can derive information. This system went on to become part of Word2Vec, a system that analyzes the relationships between words and helps to strengthen the “knowledge graph” that Google’s search engine uses to identify connections between related concepts in users’ searches. That’s just the beginning for Le. He and his colleagues at Google recently published a paper on advanced neural networks – software that’s designed to reflect the networks of neurons used by the human brain – that’s helping to advance another discipline in the field of artificial intelligence known as natural language processing, a discipline that may ultimately allow machines to understand human symbolism, subtlety and even sarcasm. Building upon Le’s word mapping, he hopes to develop systems that can translate more complex ideas into information that a computer can process in much the same way as our brains. Using things called “recurrent neural networks” – more advanced forms of traditional neural network software – machines could eventually translate entire sentences and paragraphs into numbers that it can sort, group and store. That is, they could acquire and use information just as humans do. This ability has virtually limitless potential. The world’s most powerful computers could do more...

Read More
New Brooklyn Incubator Puts Artists and Engineers Under the Same Roof
Oct27

New Brooklyn Incubator Puts Artists and Engineers Under the Same Roof

Janna Levin, a professor of astrophysics, ventures out of her university office a few times a week and heads over to Pioneer Works, a new kind of incubator in the city. Pioneer Works is owned by Dustin Yellin, a resin sculptor in the Red Hook neighborhood of Brooklyn. A few years ago, Pioneer Works was an old iron works factory that Yellin acquired and converted into a thinking space. Here, scientists and artists rub shoulders, share ideas, and inspire each other in new and creative ways. The 27,000 square foot space was purchased in 2013 for a pricey $3.6 million and has since been transformed into an open artistic thinking space that Yellin regards as his masterpiece. The integrated research space was inspired by Buckminster Fuller, Yellin’s biggest influence. Matthew Putman, a nano-microscope developer, meets with professor Levin weekly at Pioneer Works to discuss the future of his company. Although the microscopes have been in development for years, it wasn’t until Putman met Yellin and learned about his Pioneer Works space that he began to think about his project differently. When Putman opened up workshops on how to use his microscope to artists residing in the space that new ideas about his product came to the surface. Artists suggested an external screen for the microscope so that users could both capture images and use the microscope simultaneously. What seemed like a glaringly obvious addition to the microscopes functionality wasn’t apparent to Putman until he showed his design to a new group of users: artists. Levin, a black hole researcher, says that the traditional university office environment works for narrow and focused research, but to get creative she needs to get out into the world and see things from a new perspective. When thinking about black holes millions of light years away, Levin finds it easier to think about the abstract when she isn’t surrounded by results-only driven research in the university world. Although Silicon Valley isn’t yet jumping at the idea of collaborative creative and technological spaces, Yellin believes that once more people see the way his co-residents work, they’ll soon be inspired to change the way they do business...

Read More
Inside the War Room: The Competitor Chop Shop at General Motors
Oct20

Inside the War Room: The Competitor Chop Shop at General Motors

The legendary Sun Tzu once said, “It is said that if you know your enemies and know yourself, you will not imperiled in a hundred battles.” After the quality blunders of the 1970s and 1980s, General Motors has learned to listen. Just down the street from Buddy’s Pizza in Warren, Michigan is the General Motors Vehicle Engineering Center, a mechanical house of horrors where Toyotas and Mercedes are put to the knife. Approximately 100 technicians tear down competitors’ vehicles bolt by bolt, scanning each component into a 3D CAD database where the parts can undergo model animation or be manufactured by 3D rapid prototyping printers. In one sense, the chop shop is nothing new. All companies compare products.  However, GM has gone the extra mile. Before so much as a hubcap is removed, a 3D mugshot of the vehicle is scanned into a computer. This is neither a simple nor an inexpensive process. The car must be studded with self-adhesive labels, which cost $1,200 for eight rolls, and then scanned by a stereoscopic camera. Technicians use white- or blue-light image scanning for macro measurements, with each blue-light camera equipment suite sporting a price tag of $180,000. As the deconstruction continues, technicians use red-light laser scanning technology to create a full CAD model of the car, its electrical system, its powertrain, and absolutely everything else. Even the interior seam tolerances are recorded. At the end of the process, the sample vehicle will be 100% reverse-engineered. Engineers can use the treasure trove of information to decipher how Hyundai snagged an extra half inch of  head room or how Mercedes wires its windows. But General Motors goes beyond that of destroying its competitors’ vehicles. Every so often, it performs an autopsy on one of its own, old or new, to check quality. Since auto manufacturers commonly share suppliers for basic parts – bolts, sheets, etc. – then GM’s discovery of a substandard electrical component could help Ford as well. Call it Good Samaritan behavior. But at the end of the day, General Motors is working for General Motors. Immense punch presses, water-jet rigs and CNC machines hide in the back, ammunition for extensive prototyping. Because this – this is...

Read More
Throwback Thursday – 9/11/2014
Sep11

Throwback Thursday – 9/11/2014

With all of today’s modern technology, it is tough to envision a time in which prospecting for various types of ore was done with nothing more than a few wooden and brass instruments. Even more surprising is the fact that some of the most advanced prospecting gear of the 19th century was, in fact, powered much like a musical instrument. These devices, referred to as blowpipes, helped miners, excavators, prospectors, and entrepreneurs throughout the world in their search for the mother lode. Mass-produced blowpipe kits were sold to mining engineers around the turn of the 19th century, but their use in this profession extends all the way back to 1751 when it was used to discover deposits of nickel. While the tools may seem crude by today’s standards, they were actually quite ingenious, and this is one of the reasons that they remained so popular in some areas all the way up through the 1960’s. What made these testing kids so invaluable was the fact that they were portable, and this meant deposits and veins could be tested right at the source. A mining engineer used their blowpipe kit by first removing a small sample of ore and then weighing it. The ore was then placed with lead pellets on a dish in which sat the end of the blowpipe. The engineer then lit a small alcohol lamp in order to heat the ore – and this is where the ingenuity of the blowpipe came into play – by consistently exhaling, the user was able to oxygenate the flame to an extremely high level, which drove the temperatures well above 2,000 degrees Celsius. By extracting various elements and reheating the lead pellets, pure metals were separated and the process was completed. It may seem odd that this technique was taught in classes all the way through the 1960s, but that is just a testament to the ingenuity of mining engineers in previous years. With simple tool kits made from brass, wood, bone, and glass, engineers kept prospecting and mining industries running for hundreds of...

Read More
Improving an Existing Product
Aug08

Improving an Existing Product

Developing the next generation of a product can be daunting for a company already operating at full capacity.  DCI Engineering has experience analyzing existing designs to identify necessary improvements for next generation products.  The strategic prototyping process employed by DCI  minimizes the total cost of  product development by optimizing production processes for cost reduction, while maintaining high build quality.  Partnering with the experts at DCI ensures that all design and prototyping needs will be met for rapid completion of an improved next generation product. EMC Corporation is a leading global technology company that develops, delivers, and supports information infrastructure and  technologies and solutions.  They were under strict time constraints to due a predetermined global launch date.  They asked DCI to improve the design for manufacturing, reducing the unit cost, adding features, and preparing it for  scaled manufacturing.  The project was completed before the deadline, exceeding EMC Corporation’s expectations in terms of quality and cost. Having a 100% secure environment is paramount in order to protect intellectual property.  DCI works with each company to ensure that all documentation regarding ownership is filed in a timely and complete manner.  DCI operates in a high security location in Massachusetts and all information stays onsite.  You can be assured that your ideas, the minute they are share with DCI, are protected by a team of knowledgeable data security...

Read More
Prototyping Process
Aug08

Prototyping Process

Successful prototyping requires a systematic process with specific deliverables and reviews at each development phase; this enables continuous judgment and open decision making on product scope, risk, schedule and budget. The following is the description of an ideal phased prototyping process: Phase 0:  Concept and Feasibility Phase 1:  Design, Prototype Engineering, Assembly and Test Phase 2:  Engineering-controlled Production of Advanced Prototypes for Field Testing Phase 3:  Preproduction and Manufacturing Phase 4:  Sustaining Engineering Phase 0:  Concept and Feasibility In this initial phase, engineers should review the project requirements, develop a comprehensive plan of action, and integrate all key parties into the development: the Client, designers, engineers and supply chain specialists.  A specialist should run an investigation of alternative product concepts and associated manufacturing methods, followed by an evaluation of these using simulations and physical models.  Developing conceptual, mechanical and electrical layouts to identify risk areas and generate a preliminary estimated BOM (bill of material) allows experts to determine overall feasibility.  At the conclusion of Phase 0, The models and concepts are discussed in order to select a concept based on the feasibility and risk findings.  This phase is concluded with revised estimates and a project plan for subsequent stages. Phase 1:  Design, Prototype Engineering, Assembly and Test During Phase 1, The design and analysis are completed, prototype-level documentation is generated and materials are procured.  Finally a alpha prototype units are constructed for internal testing to validate function and performance according to an agreed-upon test plan.  At the conclusion of Phase 1, a comprehensive review of the prototype design and test results should be produced. Phase 2:  Engineering-controlled Production of Advanced Prototypes for Field Testing During Phase 2, final design and documentation adjustments are completed based on test results in Phase 1.  Advanced prototypes are built by skilled engineering technicians.  Extensive field-application trials are conducted jointly between engineers and the product owner.  If any issues arise during this testing, additional design changes and associated documentation updates are made.  At the conclusion of Phase 2, a comprehensive review is conducted and a production ready documentation package for manufacture should be prepared. Phase 3:  Preproduction and Manufacturing Phase 3 consists of customizing the manufacturing process and allows for the scaling up of product capacity.  Continual assessment, quantification and reports of Supply Chain effectiveness are crucial. Phase 4:  Sustaining Engineering During Phase 4, although thorough due diligence in phases 1 and 2 will reduce the chance, the customer may need ongoing engineer support due to various reasons in phases 3 and 4: (1) the product may need some modifications after actual use by a wide range of customers under diverse operating conditions that may not...

Read More