The IBM Research Laboratory in Brazil, opened in 2011, has published 412 articles in scientific journals and has submitted 250 patent applications to the US Patent and Trademark Office (USPTO), of which 40 have already been granted. The patents relate to nanotechnology, cloud computing, cognitive computing, and the internet of things, applied mainly in the oil and gas, mining, agribusiness, finance, and health industries. The North American multinational has thus established itself as one of the most productive business innovation centers in Brazil.
The company’s research and development (R&D) department has two laboratories in the country—one in São Paulo and the other in Rio de Janeiro—with 140 employees dedicated to innovation. According to Ulisses Mello, research director at IBM Brazil, the country is the group’s global leader in the development of cognitive computing solutions in the field of natural resources, and the purpose of the research laboratory, opened six years ago, is to “take advantage of Brazilian expertise in the production of agricultural products, ore, and oil in order to learn about, produce, and export technology, as well as commodities.”
In April this year, the company expanded its R&D activity with a US$4 million investment in a new nanotechnology laboratory in Rio called NanoLab. IBM Brazil has been researching nanotechnology for four years, and has set the company benchmark in microfluidic research (the study of fluids on a microscale) and the development of microdevices, similar to computer chips, for chemical analysis. The new laboratory will provide researchers with the equipment needed to identify, manipulate, and test nanoparticles, such as high-precision optical and atomic force microscopes, 3D printers, and hardware and software testing tools.
IBM’s nanotechnology and microfluidic research is focused primarily on the health, agribusiness, and oil sectors, and has resulted in 25 patent applications at the USPTO. One of the most promising areas of research is investigating how to increase oil extraction from reservoirs by analyzing, at the nanoscale, the behavior of liquid oil molecules in contact with a solid material.
The NanoLab team studies different techniques and materials that can increase oil extraction. Laboratory manager Mathias Steiner says that 60% of the oil in a reservoir remains in capillaries in the rocks, which are sometimes only tens of nanometers wide. “A production increase of just a one percent, using the results of this research, would mean nearly a million more barrels of additional oil available each day, worldwide,” says Steiner, who coordinates the research. In Brazil, a 1% increase to its daily production of 2.4 million barrels would result in an extra 8.8 million barrels per year.
Basic research is conducted alongside applied research. While investigating oil flow in nanocapillaries and studying the water-pumping pressure required to improve reservoir performance, the NanoLab uses Watson, a cloud-based cognitive services platform, to analyze polymers already on the market that can be used to extract oil from the capillaries in rocks. Cognitive systems are capable of processing unstructured and structured information, and can create probabilities and hypotheses based on the assimilated data and knowledge.
Ado Jorio, a physics professor and Dean of Research at the Federal University of Minas Gerais (UFMG), says that nanotechnology is still in its infancy, and having a laboratory like NanoLab in Brazil will generate countless research opportunities in the country. The UFMG NanoSpectroscopy Laboratory (LabNS) has a research agreement with NanoLab.
According to Jorio, there is a complementary relationship between the two research centers. LabNS has developed optical nanospectroscopy measurement equipment that is not yet commercially available and enables researchers to make optical measurements on a nanometric scale. IBM, on the other hand, is experienced in the production of nanodevices for a range of monitoring applications in agriculture and oil and gas. “We are going to combine these two technologies to look for possible nanoscale solutions to improve the energy cost of oil extraction and the development of grains in agronomy,” he says.
IBM, which is celebrating 100 years in Brazil this year, achieved global revenues of US$79.9 billion in 2016 and invested US$6 billion in R&D—the company does not disclose data on innovation expenditure and revenue in Brazil. In total, the group has 12 global laboratories, 8,000 inventors—comprising 3,000 scientists and 5,000 employees in other roles whose innovative work has generated patents—and 50,000 active patents. Part of the company’s research is conducted with academic partners. In Brazil, it has agreements with the Pontifical Catholic University of Rio de Janeiro (PUC-Rio), the Institute of Mathematics and Statistics at the University of São Paulo (IME-USP), the USP Polytechnic School, and UFMG.
One of the laboratory’s important fields of study is conversational technology. The idea is to use the Watson cognitive system to gather data, analyses, and the opinions of various specialists from a particular field of expertise published or disseminated in a range of formats (text, video, and images), and then offer users an organized set of information via an application. The conversational application, still under development, is known as Agria. A similar solution is also being created for the financial sector, called Finch. “Even a small-scale investor will have access to several analysts before making a decision on their investment,” says Ulisses Mello.
Brazilian innovations in cognitive computing encouraged IBM to establish a cooperative agreement with FAPESP through the Research Partnership for Technological Innovation program (PITE). IBM and FAPESP will each invest US$250,000 over 10 years to fund joint research conducted by universities in São Paulo and IBM’s Brazilian R&D center.
A first public call for projects has already been made, and seven proposals were approved in March this year. These include projects that aim to automatically estimate age by voice signal processing, studies on advanced machine learning, development of inference and learning techniques for probabilistic logic programs, descriptions and solutions for spatial games, and language processing with mirror neurons.
Watson cognitive services have been used in the Brazilian financial sector since 2015, when bank Bradesco adopted the system for its internal call center. First, IBM’s Brazilian laboratory had to enter the Portuguese language, with all its regional nuances, onto Watson. The bank then fed its data on work routines and Bradesco products into the system, enabling it to produce answers to more than 200,000 questions about 59 bank products and services. In 2016, Banco do Brasil began using the system for its customers online, where 60% of the bank’s transactions are performed. Customers can perform 144 different transactions using the system.
Watson is also being utilized in healthcare. An initiative is under development at the Mãe de Deus Cancer Hospital in Porto Alegre, Rio Grande do Sul. Márcia Ito, an IBM scientist working in the health sector, says that the Watson for Oncology solution will allow physicians to enter clinical information about their patients on the system, such as medical history and exam results, and the program will present treatment options based on global scientific evidence. “It is a clinical decision support system. But the physician will always have the final say,” says Márcia.
For physicians, one advantage of the system is that it provides up-to-date information on new scientific findings. In oncology, staying on top of the latest knowledge is especially difficult, with around 50,000 studies on cancer published worldwide every year. The Brazilian team is currently adapting Watson to the clinical language and terminology used in Brazil, as well as local clinical protocols and the regulations established by the Brazilian Health Regulatory Agency (ANVISA).
Focus on agribusiness
The importance of agribusiness to the Brazilian economy has led IBM to look for solutions in the sector. The result is open innovation platform IBM AgriTech, which uses the company’s technology, such as cognitive computing, the internet of things, and blockchain (a record of digital transactions with guaranteed authenticity), to develop local solutions in Brazil with global marketing potential. The projects currently under development in the laboratory include precision agriculture models based on weather forecasting, which allows partner companies to create tools that use this information in the field, offering guidance on the best time and day for planting, harvesting, irrigating, and applying pesticides.
Another field of research is cognitive image recognition, enabling the analysis of photos taken in the field by AI drones. The objective is to design tools that detect the presence of pests, for example, or software that counts the fruit in an orchard, or the number of trees in a eucalyptus forest for a pulp and paper company. Recently launched IBM AgriTech was developed in partnership with agribusiness companies and organizations that operate in the Brazilian agricultural technology sector.
GIRO, R. et al. Adsorption energy as a metric for wettability at the nanoscale. Scientific Reports. Apr. 11, 2017.