Machine Learning

imotion with nam nguyen used eye tracking in simulation, military, gaming, and automotive. imotion task sensor data and creates visualizations gaining understanding of muscle tensions. iMotion is capable to eye tracking and facial expression analysis. Additionally they can gather information from sensors like eeg, ecg, emg, and respiration. imotion does eye tracking in store, mobile, automotive, simulations, and vr gaming. Data analysis includes gaze replays, heatmaps, and areas of interest.

How much should you borrow? None is the correct answer. If you don’t have money then don’t go to college. Save your money then go to college. If you yield to temptation and accept a loan then your loan payments once you graduate shouldn’t exceed 10% of your expected monthly gross. You will have 10 years to repay the student loan. If you earn $40,000 then your payback will be 4,000 a year (10%) or about 340 dollars a month for a $26,000 dollar loan. The average undergraduate loan is about 26k. 50 percent of the student enrolled in college do not graduate and leave the college with heavy student debt. If you graduate making less than 40k the loan repayment amount will be overextended in ratio to earnings. Instead of 10 percent of your wages the loan will represent 15 to 20 percent. The heavier debt loan increases the risk of default.

This is not financial armeggedon for you. Your not under reckoning judgment. Start fixing your financial problems. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Thoughts on electric trucks and deep learning networks --- Send in a voice message: https://anchor.fm/david-nishimoto/message

xcient is the worlds first heavey fuel cell truck being mass produced. Xcient is a 190 kw fuel cell system. In commericial use in Europe. Switzerland is receiving 10 xcient trucks per month. The fuel cell is a present day reality. For large semi haul, hydrogen is the only alternative above 34 tons. The goal is 1600 units by 2025. Hyundai has deployed 9000 fuel cell cars. Germany is creating a hydrogen refueling road map. Hyundai target is 125,000 fuel cell cars by 2025. The cost of a fuel cell vehicle will be comparable in price to electric by 2025. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Hydrogen has the power density to create a supersonic electric plane --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Riffyn solves the data fragmentation problem and helping to make data more effective. Data scientist spend about 80% of their time cleaning up data and not discovering pattern. Data Scientist spend alot of time analyzing spreadsheet trying to identify what the features mean. The cost of research and development is staggering. We don't know alot of about biology and chemistry and getting more data will improve understanding. Riffyn finds unexpected relationships in the data. Riffyn uncovers defects in data discovery. It takes about 10 years for a chemist to become proficient in chemistry. Machine learning can speed up the learning and proficiency. Noise in the data causes missed opportunities. Riffyn operates in recipe driven environments. For a process, Riffy creates data collection steps and joins the data between the steps. The data is annotated and tagged for each step. The user can then download the collected data into a csv file. Fit to least squares is used to determine maximum yield. https://lnkd.in/g

Many people believe a house is an investment and safer and more profitable than stocks. A house is not a profit center, but a cost center. Each month money must be spent on a house. A house produces nothing; it has no revenues; and no profits are earned. You never hear of a factory being built with increases in house prices. Nor can housing be exported reducing the trade deficit. However, much of the meager savings and massive borrowing is going into housing. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Ai for business process --- Send in a voice message: https://anchor.fm/david-nishimoto/message

cusp can autonomously7 patrol regions around the navy ships and protect them from smaller boats --- Send in a voice message: https://anchor.fm/david-nishimoto/message

unmanned cargo ships will be navigate the oceans. The unmanned cargo ship will combine the technology of unmanned drones to control the ship. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

chorus.ai is customer centric. What can we learn from our products? chorus captures interactions in phone and email. chorus.ai stack includes voice recognition. chorus analyzes conversations. Three reasons to analyze conversation: differentiation the closers, cut down on ramp time, and reducing the 80% of calls lost. chorus.ai decompose conversation in semantics and core building blocks. The data is used to determine risks for the deal and topics and themes from conversations. Performance must be collaborative and friendly. Conversation must be open ended to discover the pain the customer is experiencing. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Make debt painful --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Debt is simply borrowing money. Saving is the new maximum. Spend what you earn and no more. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

verge genomics is using ai to discover drug for health. The power of genomics lead to the creation of the company. the goal is to build a scalable drug discoverable engine across every human disease personalized for each patient. verge is identifying the exact gene signatures causing disease.  Using biological computation prediction into the lab. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

in one case study icertis ai automated 70,000 agreements representing legacy contracts in a data repository. Icertis au helped the company to get information out of the documents. Contract generation manually done is much slower process. icertis helps locate different documents and work with them in a efficient way. Icertis analyzes risks and ensures compliance --- Send in a voice message: https://anchor.fm/david-nishimoto/message

forethought helps make everyone a genius at their job. can we use ai to help people with work. Everyone has the information problem. go to market. customer support is an important part of the customer experience. automating tasks and assisting employees answer questions and write code. ai can be a companion or tool for all knowledge work. ai took a leap forward in 2017 when forethought started. squad dataset from stanford helped nlp with question and answer. nlp is starting to industrialize. BERT was a state of the art NLP model capable of human language natural tasks. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

Central bank policies are the battle that hedge funds capitalize on --- Send in a voice message: https://anchor.fm/david-nishimoto/message

atomwise origins is chess Computation biology design of a pharma molecule. Machine learning and big data sets and biology kept creating molecules in simulation that worked. For a drug to work it needs to stick to a protein and arrest the function of the protein. Other proteins need to continue to function properly with the molecule.  The problem becomes a prediction problem. Does the molecule stick, yes or no? test the molecule will bind to the protein. the drug molecule blocks up the protein. Proteins are 3d matrix: oxygen, carbon, hydrogen extra. The 3d grid is feed to the deep learning network cnn. institutions with specific proteins can have a list of possible molecules that will bind to it. atomwise does the computations and sends the results to the partner institution which physically tests the molecule list and sends atomwise back the results. --- Send in a voice message: https://anchor.fm/david-nishimoto/message

AIG runs out of money and the government must provide $170 billion in bailout funds to prevent bankruptcy and owns over 80 percent of AIG. --- Send in a voice message: https://anchor.fm/david-nishimoto/message