Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API

From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Search - "natural language processing"
-
I am traveling 550 Km (9hrs) just to give my first interview for the position of Jr. Natural Language Processing Engineer.
Wish me luck...8 -
I'm looking forward to natural language programming.
The ability to code by explaining what you want to happen and having a neural network work out the fine details in an optimal fashion with evolutionary techniques.
I look forward to the super AI. I don't think they will necessarily be evil, however above a certain point we would seem like ants to them... And when was the last time you checked if there was an ant where you were to put your foot? It's not malicious... It's just not worth your or their time.29 -
During one of our visits at Konza City, Machakos county in Kenya, my team and I encountered a big problem accessing to viable water. Most times we enquired for water, we were handed a bottle of bought water. This for a day or few days would be affordable for some, but for a lifetime of a middle income person, it will be way too much expensive. Of ten people we encountered 8 complained of a proper mechanism to access to viable water. This to us was a very demanding problem, that needed to be sorted out immediately. Majority of the people were unable to conduct income generating activities such as farming because of the nature of the kind of water and its scarcity as well.
Such a scenario demands for an immediate way to solve this problem. Various ways have been put into practice to ensure sustainability of water conservation and management. However most of them have been futile on the aspect of sustainability. As part of our research we also considered to check out of the formal mechanisms put in place to ensure proper acquisition of water, and one of them we saw was tree planting, which was not sustainable at all, also some few piped water was being transported very long distances from the destinations, this however did not solve the immediate needs of the people.We found out that the area has a large body mass of salty water which was not viable for them to conduct any constructive activity. This was hint enough to help us find a way to curb this demanding challenge. Presence of salty water was the first step of our solution.
SOLUTION
We came up with an IOT based system to help curb this problem. Our system entails purification of the salty water through electrolysis, the device is places at an area where the body mass of water is located, it drills for a suitable depth and allow the salty water to flow into it. Various sets of tanks and valves are situated next to it, these tanks acts as to contain the salty water temporarily. A high power source is then connected to each tank, this enable the separation of Chlorine ions from Hydrogen Ions by electrolysis through electrolysis, salt is then separated and allowed to flow from the lower chamber of the tanks, allowing clean water to from to the preceding tanks, the preceding tanks contains various chemicals to remove any remaining impurities. The whole entire process is managed by the action of sensors. Water alkalinity, turbidity and ph are monitored and relayed onto a mobile phone, this then follows a predictive analysis of the data history stored then makes up a decision to increase flow of water in the valves or to decrease its flow. This being a hot prone area, we opted to maximize harnessing of power through solar power, this power availability is almost perfect to provide us with at least 440V constant supply to facilitate faster electrolysis of the salty water.
Being a drought prone area, it was key that the outlet water should be cold and comfortable for consumers to use, so we also coupled our output chamber with cooling tanks, these tanks are managed via our mobile application, the information relayed from it in terms of temperature and humidity are sent to it. This information is key in helping us produce water at optimum states, enabling us to fully manage supply and input of the water from the water bodies.
By the use of natural language processing, we are able to automatically control flow and feeing of the valves to and fro using Voice, one could say “The output water is too hot”, and the system would respond by increasing the speed of the fans and making the tanks provide very cold water. Additional to this system, we have prepared short video tutorials and documents enlighting people on how to conserve water and maintain the optimum state of the green economy.
IBM/OPEN SOURCE TECHNOLOGIES
For a start, we have implemented our project using esp8266 microcontrollers, sensors, transducers and low payload containers to demonstrate our project. Previously we have used Google’s firebase cloud platform to ensure realtimeness of data to-and-fro relay to the mobile. This has proven workable for most cases, whether on a small scale or large scale, however we meet challenges such as change in the fingerprint keys that renders our device not workable, we intend to overcome this problem by moving to IBM bluemix platform.
We use C++ Programming language for our microcontrollers and sensor communication, in some cases we use Python programming language to process neuro-networks for our microcontrollers.
Any feedback conserning this project please?8 -
Started new course called "Introduction to natural language processing" in uni. I am super bad at doing regular expressions and don't understand anything about them.
Saw the first weeks homework. Have to do i.e. some text cleanup with regex... I was sad. But now after reading the course material and trying some of the exercises I'm super excited since I'm actually doing something "real" with it.
Do you guys just love it when teaching material is well written? I do.3 -
I just wanted to try out apache open natural language processing. But why is building java from source such a pain? Totaly unintuitive. 😕
-
What's the best natural language processing software that won't f you up?
I'm a big fan of Alexa's capabilities but we all know that Alexa is to security what North Korea is to democracy.
Is there any software that can compete with powerhouses that are Alexa, Google home, Siri or cortana?4 -
Built a pretty slick chat bot for my company’s conferences that used Google’s Dialogflow for natural language processing and conversation state
It worked from a web chat or SMS. Allowed manual responding by agents as well as the chat bot. Pulled dynamic answers through a 3rd party API integration
Most common questions “what is the wifi password” and “tell me a joke”
Project was killed after 2 conferences - thankfully it only took me a few weeks to build4 -
!rant
That moment of satisfaction when you finally complete a report on Natural Language processing by reffering 6 IEEE paper and 7 presentation document within a night before submission. -
I feel like I may be done with dev... the imposter syndrome has been hitting hard lately. I really want to get into Natural Language Processing, I'm currently looking at skip-gram parsing a dictionary using Word2Vec, then I came across a paper called dic2vec which looks promising.. half the time, I just think I'm barking up the wrong tree, or that it's been done before. Most times I conclude that I have nothing new to offer and there's gotta be half a thousand people like me, striving in the same space. Possibly failing. Don't get me wrong, the state of consumer software at the moment NECESSITATES my involvement (I'm looking at you (epic games, windows) , every which way I look at it. I just don't know where or how to get going. Viva la revolution. A toast, to shitty software and exceptionally low moral *klink*6
-
Currently working on a distributional semantics database or something like that. The goal is to read natural language texts and construct a graph of words, with directed edges containing the relative distances between them. Then I want to enter one or more words and find all of the possible words which could be used before, after or in between those words, simply based on the previously learned texts. Then I want to find words which are used in similar contexts and build a kind of dictionary that way.
The end goal is to use this to define software or other states and procedures, using natural language.
What do you think of this idea? Can you imagine it being feasable for its purpose? What are your thoughts?3 -
OK, I think Ive figured out how I want to implement the core of my game.
Basic idea. Everything is made of sigils. What is a sigil? Lorewise, a representation of a concrete idea. Mechanically, it's a container for one or more ingame effects.
Again, anything you can think of, be it mundane or arcane, is a sigil. Every object attribute and method that causes a mutation of the game world state is composed of either a single or a sequence of sigils, known as a signa. You can do abstraction and have one thing encapsulate others, all that good stuff.
We go turnbased because tabletop rules. In your turn, you choose what to do with the sigils available to you. Common stuff like walk, talk, crouch, attack, et cetera are available to everyone.
You choose which effects to use, keeping in mind that order matters, and these effects either clash against or boost each other according to internal rules. Any relevant environmental modifiers around you or status effects apply here as well. Theyre also made of sigils, so theyre just implicitly components of the spell.
Finally, we get the result. An aura. It is simply the aggregate of all the sigils, containing the final transaction that will be applied to the game world, with dicerolls, if any are needed, already performed.
So here's the crucial bit. We record the aura for each turn, taken by either the player or the game itself, which includes the sequence of sigils used. Because each of these are words, or a series of words, it can be used narratively.
Let me give you an example. You are in a swamp, you use your turn to move around. This gives you an aura like so.
src: player, location: swamp, status: ok, walk.
We push that to a stack the game can see.
Now the game has to take its turn, and will look at the stack to get ideas. So lets say this swamp hindered your movement, gave you some trouble getting around. This is noteworthy enough as it affected something in the game world, and theres nothing else going on, so its made into a topic. The game will now pay attention to it and make it available for talking about.
topic: swamp-slow on player-party.
Everything is sigils, remember? Speaking is an effect, as is giving you something to talk about. Adding a topic results in an aura, push that to the stack.
Maybe then the game decides have the players companion make a comment on this topic. OK, just pair speak with other sigils to make an expression. The current topic is simply about a single negative effect, so one or two should be enough for a comment.
Factoring in the companions traits, also sigils, and the fact that for now this only a minor annoyance, we end up with an aura.
speak: topic -> complain about.
Do you see where this is going yet? The spellmaker gives us the means for text generation. The key here is instead of processing natural language like a bunch of dumbasses we just have an array of structures meant specifically for this purpose.
Anyway, the player now takes their turn. They can keep silent and continue moving, or say something as well. So let's say you pick the speak sigil too because why not, the current topic is also available to you, so you can reply. Depending on what the last message (aura) was, you get a range of options, and why not, let's factor in the player's chosen traits as well.
speak: shutdown -> complain about topic DOT change-topic -> mcguffin-quest.
We're on a mission to recover some plot stuff so shut up.
This results in further state changes, albeit behind the scenes. Our companion isn't happy, and we have a sequence of structures to prove it. Actions taken in relation to a topic can be associated with it if they're deemed significant, id est they caused a mutation of the game world's state.
speak: remind -> topic-name.
Thats right. Companion dude WILL remember that, you son of a bitch, and you bet your ass hes gonna bring it up if you keep acting like youre his boss or some shit. You dont even pay him!
You may say its such a small thing, but it all matters in a story. So we keep track of it, see what we can make this branch off into once a little more time has gone by.
What do you mean crazy, Im a fucking patrician of tenebrous arcana. You solve the problem by not solving the problem, obviously. Fuck natural language processing, all my homies use Litechnics to make narrative-driven programmable representations of the human condition.
Further rules include breaking gameplay down into scenes and acts as if it were theatre, because why not, and having baddies cooking shit up behind the scenes. Mechanism for that is the same, read the current stack and output structured wordsoup based on it.
We can even have a chorus narrating.
Or an audience reacting lmao.
OH GOD.
BUT I'M SO FRIGGIN SLOW OF A DEV.
WHAT IF I DON'T LIVE LONG ENOUGH TO IMPLEMENT THIS?
SHIT.
WORST SHOULD COME TO PASS,
PLEASE AVENGE MY DEATH.5 -
What's the best way for a beginner to learn Natural Language processing? Don't want to do a coursera or udacity course. Just a simple tutorial would help. Any suggestions?7
-
Hey guys, wondering if anyone knows of somewhere in Melbourne Australia that is hiring interns/Students.
I am a full stack developer under nodeJS/javascript.
I have been using python for the last 5 years (2 and 3).
In my spare time I work on computer-vision and natural Language Processing.
I am looking for any jobs around my area.
I will leave my GitHub here incase you would like to look.
Https://github.com/crazywolf132