authors, illustrators, graphic novelist is, al of these different authors there all day, over 100,000 people come in and celebrate books and reading. you can't have a better time, i think, and i'm a little prejudiced because i'm a librarian but any read her, a book festival is the place. the communicators recently visit microsoft's washington, dc headquarters to hear about

their lobbying efforts and to learn what is happening in microsoft's research labs. >> now in the communicator wes want to introduce you to fred hem fry, the vice president for government afared for the microsoft corporation. mr. humphrey, with that title, what your duty? i'm the ambassador for microsoft when it comes to public policy, innovation and technology. grew up on capitol hill and participated in community activities when it comes to talking about technology. >> what are some of the issues where microsoft was intersecting with public policy. >> guest: we are intersecting when it comes to artificial intelligence, when you think about those policy issues, and think about vehicles or research and development, high-skilled immigration, take note of different individual participating today and part of

our microsoft research. a lot of them are high-skilled visa holders who are making a difference and making many contributions to not only the u.s. but to the world when it comes to innovation and technology. >> host: sounds like one of the new issues on the communicators is artificial intelligence, we have not talked about it much yet. where is the public policy. >> guest: a.i. is in early stages. frankly, it's kind of like what think about one of our main goals is freedom to innovate. so when we think about ai, it's important that wet get our arms around and it learn more. what educating members of congress at the federal level, state and local level so they have an understanding and get to see how a.i. can make a difference. >> host: recently net neutrality has been in the news. what's microsoft's position?

>> guest: we like what is taking displays where the laws are as a result of what took place with the -- under chairman wheeler at the fcc. we're pro and for net neutrality. >> host: it's going away as title 2 function. >> guest: yes, and we have concerns as well as many other in the industry you'll see a lot of different associations that tech companies are part of to focus on that area. we plan to educate as well. >> host: privacy issues are in the news. where is microsoft at. >> guest: striking the right balance. in some cases we look at lawful access issues, making sure that the consumer's data is protected and we have rights and reasonable rules when it comes to privacy protection. >> host: now, fred humphrey, what is your background? >> guest: i've been fortunate enough to work at the federal level.

worked on capitol hill. i was on he hill for u.s. worked for governor mcwhorter, i work on the legislative staff and worked for a member of congress, bishop. worked for another company. one time i did advocacy and was a lobbyist so i have had opportunity to work on campaigns. so i worked for a governor, and now i have the pleasure to work for a company that i really think is going to make many contributions to society. >> host: the last couple years microsoft seems seems to be res. >> guest: we have always been kind of flying high and doing out standing thing. it's an exciting tomb under our leader, outstanding ceo and when it comes to the cloud, the

intelligent edge, and our mission is to empower every person and organization on the 2013 achieve more. think you'll find today as you go the different displays -- we have 12 here. a microsoft resurgence from 1991. and what you are seeing on the drones and efficiency that make a difference for farmers and using tv white spaces, which is something that the fcc is looking at. helping farmers be more efficient and effective and as a result be able to do things cheaper. you see the hollow graph and -- holograph and augmented reality. a lot of things when you walk around to have a societal impact and have -- relates to the cloud, relates to efficiency, relates to a.i. and the internet of things.

>> host: it seems there's a big investment by microsoft in a.i., a lot of the plays about artificial intelligence or have that come opinion -- component. >> guest: we have been focused on it for a while. you'll have the opportunity to meet eric, the leader in a.i., technical fellow. it's the future. that what you'll see here today, future technology that on the horizon. >> host: we will walk around in a minute. one more issue. cyber security. what would you like to see congress do? >> guest: on the cyber side, we really need a lot of work on just what is the rules of the road, the framework on how we look at cyber security. i actually think that congress, whether it's bipartisan support on that, i think that the trump administration is looking to focus on that as well and we

look to participate. one thing that is very important is we have different cyber attacks, different cyber issues when it comes to, i'll say, viruses or malicious software, always important. but when you get a popup that says you need to do a update, it's very, very critical for al of us. as well for companies to make sure they're doing their part to protect their software and technology. >> host: when brad smith, the president of microsoft, sent out a blog post talking about government should not bestockpiling vulnerabilities, what did he mean by that. >> guest: i think that speaks for itself. everybody needs to kind of be responsible when it comes to cyber aspects, and when things

unfortunately get -- are available on the web, you have some bad actors out there and you need to think that what you're keeping and needs to be protected. just think that on that blog, on the stock puhling, kind of -- stockpiling, speaks for itself. >> host: fred humphrey, you're one of the big five, big six tech companies. do you work with the other big tech companies, washington offices, on a regular basis. >> guest: absolutely. not just the big tech companies. we work with many different -- microsoft has a lot of partners. it's a big ecosystem. we work with the big tech companies and work with them on a.i., privacy, cyber security, from autonomous vehicles, drones, whatever it may be, the different public policy issues. we're constantly working with

them, and still the protection of intellectual property, trade, and those type of issues. so, there is a -- what you find is there are lot of public policy agreements when it comes to the tech industry. then on the business side we all do what we need to do to be competitive and to advance and expose people to our technology. there's really a lot of harmony mook the companies when it -- among the companies when it comes to public policy. >> host: how big i microsoft? how many employee snooze i think globally it's 118,000 employees. >> host: fred humphries is vice president of government affairs at microsoft. thank you for your time. >> guest: thank you. >> the communicators is at the washington, dc office of microsoft for their annual tech fair and now we're at the farm beats booth. mr. chandra, what are we looking

at. >> a project called farm meat, and the goal of the project is to solve the world residents food problem. the world residents food production needs to double by 2050 to feed the population. how do you do that? the moisture levels are shrinking we are looking at the solution for farm beef to previously people have been becoming an this problem and have proposed a few solutions. these techniques have known to improve these. for example, the idea of -- instead of treating a farm as homogenous, you treat it as let the genius, apply water when and where it's needer, first lies enwhen it's need -- fertilizer when it's needed and plant closer together. >> what does artificial intelligence and i.t. have to do

with agriculture? how do you use those. >> guest: we need these kind of treatments you see here. we need to be able to build dense water maps, like how does the nutrient move. how do you build these kind of maps? existing approaches to build this irvery expensive. cost dish was an expo and the cheaper sensors were $8,000 so that's tattoo much for a farmer. the goal is to reduce the cost offing a ashing a agriculture. the first one is about connectivity. most of the farms are in at the middle of nowhere. so the reason the sensors are so expensive they use cellular

solution or satellite but that's expensive. we use the td wide spaces, pose you have a wi-fi outlet at home that you plug into your house, if you can access it's few while away? we send the i wiaa identify sigl in unused tv channeled. on channel 9 we can send the wi-fi signals without interfering with the tv transmissions on channel 8 or 9. tv frequencies and the lower frequencies have the same power level as wi-fi. the signals can go four times further. and this is increased speaks. to crops, canopies, signals can go on farther. the interesting thing about agriculture is in the farms away from the cities there are lot

more empty tv channels. and this tv -- ick you have 40 tv channels can talking -- so just like wi-fi connect your house, this can connect your farm. >> host: what are the thing this farmer can do. >> guest: the other thing is the emerging -- machine learning. you don't need to -- instead of having -- usually you have a feed map and using this approach you can have a few sensors, merge with the drawing, and we're doing everything that is running there and is running on a pc in the farmer's house. don't need to ship the data from the farm to the house. we're in deep learning at the

gateway, this is actually trained on companies but you sere seeing the pc right there it's not in the cloud. all local compute -- cute addition so we bringing -- compute addition so we are bringing down the cost of the system. you're doing machine computer vision, machine learning, deep learning. so the kind of things we are de -- >> host: why is it important that we know that these are people, somebody is wearing a tie somebody hat a hand bag and cows are running around. >> the context of cows, if you have them in barns you, know if there's a intruder, some i cows counseling and. this gist for demo purposes.

>> going back to the deployment here, quickly walk you through the kind of scenarios we can enable. so these are some of the deployment. see the deployment in car nation, essex. the kind of insight. this is a four kilometer stretch. the companies are in past steur and these are wait where the cows are pooping around and this is a stray cow needed to be herded in. we can flag these moisture maps, we could flag this part is miss without having a sensor here. we are able to come up with these kind of sensor heat maps. we're able to flag parts of the farm that the ph is not correct. the farmers started using this storage monitoring. have the door open so the food

product can go bad and there are many more scenarios that we are running. >> host: are these being used in the real world today? >> guest: yes. we are working with farmers, with partners on getting this deployed. there are two live deprime minister wes have, one in car nation, one in upstate new york, one in india. we are talking to many more people. this is a complete system. ingest data from sensors and cameras and provide useful services in the end. what services we provide is specific to the farmer. some one wants to do cow monitoring, another one might want position maps and we work with them. >> host: what's your background? do you have a farming background or software. >> guest: i'm a ph.d in computer science from cornell but i did degree up on farm growing up so growing up in india, every summer -- four months a year i would spend on

the farm in india. this is sugar cane farming, wheat, rice, and eventually we can benefit the farmers that grow food for us, make them more profitable. >> this the farm beats display at the microsoft tech fair. >> and jamie burgess, what is cognitive service snooze cognitive services is our collection of microsoft brains that sit in the cloud and can do specific tasks, things like text, face detection, vision. if you look over here at the services web site, there's a list of all these services we released, even in preview. some i'm really proud of. we can see here that there's a group of vision -- you can use

motion detection, speech, text, all these kind of things. at the end of it, people have now -- every time now with microsoft when you see texts, the language is irrelevant because we can translate to any language. so, this is a real -- >> host: now, microsoft as your available to anybody? >> guest: yes. so, microsoft is based on -- all encompassing technology for cloud, ands aure is where we can call response these -- these are very easy to set up, very easy to use. i'll give you an example of one in just a second. they're available on a public web site. you can try these easily with an outlook account, for example. here's a very simple example of one of our cognitive services. i want to search for things that

are happening in dc so i look for washington, dc. i found a can be that is publicly available, faq on the city. pulls up this web site. have i ever read an faq? probably has many is a have manuals which i grand total of one so i grab that url and look in and put in the url you can see it's captured -- any format, captured the questions and the answers. all the questions and answers and this is what it would look like if i published it to a person's web site. can ask the question, and it will go off and found the answer and present it to me just like that. the whole process of grabbing the web site or uploading a document and producing the q bea is two minutes. now, web sites like that, the

benefit of technology, and they could actually start trimming down the question. for example, what time is the white house open? any simple example. >> host: is there artificial intelligence in what you just did. >> guest: a little bit of a.i., more towards the machine loading i. can add other ways to invoke a question. so my question i was, ick tour the white house? i could answer, would it be possible for know wander robbed the white house? the different context i would ask the question. that's where the machine learning comes in, to offer the same solution to -- >> host: let's work down and see what else microsoft has in the way of cognitive services. >> guest: another example is this video indexing service. a useful technology if you're creating a video to deliver to the masses, whether it be a stage presentation or a discussion or -- this is our

illustrious leader. our public web site, they're yet to be fully released. actually in -- still sharing them out to the public to learn from them. this is recognizing faces from the video, also recognizing the sentiment on which -- so we see how they deliver it, and we are showing scott guthrie and you seek the dna looking process, when scott is actually talking. from an accessibility standpoint, when i want to create the videos i also want to create transcription so people could be hard of hearing of another language. so it's 50 minutes to go through the transcription process and it's in text. whenever we have text, you can

think of it as language irrelevant because i can check on any language. >> y to go back, you have had gentleman up over here, mr. humphries, -- guthrie. what was his point if he was in the speaker. his reaction? no one of the speakers on the group of speakers we are talking about. >> host: is this connected to what we're doing up here. >> no. that's just the conversation around 2017. >> host: okay. >> guest: a little longer to reload. >> host: all right. so this is -- >> guest: let's say what if i was superman? get ready to start in the middle.

the interesting thing is, the bot behind looks look an emotion, everyone is laughing and then if everyones frowny, the book gets upset you don't like -- >> host: all right. >> guest: the technology behind that is impress. knows who superman and is knows taking a photo on your face and then imposing it over the face. >> host: you have your age detector. >> guest: notice how it's completely anonymous, no names or anything it i can trade the mod model to add -- like this chap here, harrison ford, and i could even add myself in there as well and knows how. because i trained it. >> host: it knows your name. >> guest: yes. added it. >> host: makes people older than they are. >> guest: sometimes, depending on the lights. the interesting thing is

people's sentiment change based on the age portrays. you've get a low number, everybody is happy. in in the real world you analyze crowds if you were having aned a vert in a shopfront and wanted to capture people honestly, you could do. >> host: a privacy issue? >> guest: privacy is one of our most important parts. as is security. you can fool it with a piece of paper but we would only use this where they're proper signage and being informed that this is captured but it's anonymous. the technology here is capturing -- human facers are very easy to detect. it's sending this little bit of information on a round-trip to the crowd, anonymously, and gathering the information on my sentiment, my age, and if it knows where i am. >> host: so this information that you and i are just standing in front of this, this is going out into the world.

>> guest: not the world. the cloud. you can -- >> host: away from here. >> guest: going my setup and my environment. i'm not storing it. just portraying anonymous data. >> host: when we were on here there was a percentage next to your name. >> guest: yes. that's how accurately it portrayed my age. or even my name as well. so based -- the percentage is how close to the thing it's getting. it's a common standard with machine learning. you want to know what it's trying to identify and how accurately it can identify that thing 'when i shore harrison ford, 68%, it is sure it's harrison ford. >> host: okay. again, what is the use of this. >> guest: imagine i was delivering a speech or talking to many people and i want to analyze the sentiment.

if i have a new concept i want to land to an audience talk about the concept, during the moment it's very difficult to get a feel. you try to deliverrure message so what you can do is load the video and see the speaker and not the audience. we have a new technology in one of our public places so people can vote, plus or moon news, like emojis. so i can face this into an audience and pick up their sendments. if i could talk to them i could essentially have someone analyzing the sentiment and advise me what to talk about if i'm losing the audience and need to gain their interest again. >> host: is political speech elements. >> guest: i could be doing a speech, ted talk, anything where there were many people in the audience. >> host: jamie burgess of microsoft, this is a look at their cognitive services. thank you for your time.

>> guest: no problem. thank you. >> kirk hunter of microsoft, what are we seeing. >> 3d printing. windows 10 has built in support for 3d printing so we're enabling the ecosystem much in the same we way have in other computer peripherals like rig printing. companies that build 3d printers can plug into windows and guys that are building applications, can easily add that their application. >> host: through microsoft software. >> guest: through the windows platform, yes. >> host: 3d print is a couple of years old, pretty mainstream. >> guest: sure. it's fairly mainstream, but it's becoming more and more accessible as the technology gets lower price point so people can get entry into using it, and then there's innovation on the high enas well. >> host: do you know what printer like this would cost today or like down here at the other end?

>> guest: the types of printers, we call it's desk top printer, range in price from, say, $3,000 to $4,000 down to $200. >> host: they've come down in price. >> guest: absolutely. >> host: nat isor job? >> guest: sometime a practice manager and my team is responsible for building all of the technology into windows to support the 3d printing. >> host: what are we looking at here on the shelf? >> guest: a wide variety of different examples of things done with printing, some done locally. you can pull that out. that's three parts. >> host: oh. >> guest: illustrating different concepts. three different ways in which you can get content for 3d printing. one is you can find models online. there's a lot of different sharing types of models. these models are much more clever than me, actually, designed that. another way you can create concepts is to scan something, so you can take a 3d scanning device and scan somebody and

create a digital model. and then a third way is actually design yourself. there are variety of dust software tools that allow you design 3d models, ranging from entry level, basic tools, up to high-end professional computer added -- >> host: are we to the point where people are putting 3d printers in their homes much as we did laser printers? sunny think so. depends on your needs. think of a 3d printer as akin to a shop tool. you had the need for a table saw because you like to work with wood, similar thing. >> host: what is your background. >> guest: i'm a software engineer by trade, and i've been at microsoft for 20 years. and just think of this as a fascinating technology. >> host: is this -- how user friendly is this what we see on the screen here. >> guest: this