Showing posts with label Artificial Intelligence. Show all posts
Showing posts with label Artificial Intelligence. Show all posts

Ubisoft's SAM is ready to be available for personal gaming assistance in service

It's no longer a wonder miracle to any users these days that technological implications is going beyond its abilities through prudences in developments making us lighter , sleakier and even easier for users to interact and communicate through modern means effectively. But in the recent steam of gaming lots of improvements has been taken place from building flexible graphic cards to tobii's eye  tracking sensory intelligence has heaps around our movements in service which tracks our eye sensory perception 

Ubisoft is been known for its video games publishing for several acclaimed and most played games like Assassin's creed  , Watch dogs, Far cry , The division , Crew and now they are onto artificial intelligence adaptation into heavy bursts for pullers and video gamers that uses Google cloud NLP platform and also Dialogflow to get enabled. It easily interprets the best user inputs and crafts it very effectively through mobile app called SAM

Infographic of SAM

( Source : Ubisoft SAM )

To get Ubisoft's SAM activated, you first need to download the Ubisoft club app for mobile and get signed up. Once user authentication process is over, the information which are associated with Ubisoft's game plays and tips are relatively indexed and will be assessed and the tips,news and offers will be sent to every user who registered it will be offered with great source assistance like Siri for Apple , Cortana for Windows , Google Voice assistant for Google and immediately service will be put into active mode

ubisoft club sam chat

ubisoftclub_sam_weather_aco_screenshot

SAM has been tested with around 400,000 questions and personally symbolize , analyze and also share news of vital information which related to every personal user's strength , weakness , agility , power factors , favorite gear and game statistics. Suppose for an instance if you are playing a game of Assassins creed and there you could not able to catch up or could not able to win any particular level or if you win rate is low, your weak points will be assessed and SAM will hand over the community related content and also pushes the user with notifications as tips to strategically beat the enemies through the agility or user's data tantalization in game while playing

So, the SAM does this task more effectively and through NLP and not only that but also every user can ask questions to SAM and also could able to vote for content updation in community participation. With it's easy to use mobility, users can engage more and compete their score cards and performance statistics anytime they want and correct it , updates it and propose new content in the app and also it gets feedback from users. 

Corti :The real-time ai co-pilot that helps to analyze symptoms and alerts emergency in fast pace

Medical necessity has grown deep and far stretched these days for everyone at any plotical conditions that we never expect to happen. When the heads of innovations advancing in one side , the other side diseases and emergency requirements for an human being gets more close with various fatalities incurred in modern era. 

 I was talking about AI in the medical stream through various ways and output efficiency results seems to be improving than earlier . Earlier in few of my posts , i was telling how artificial intelligence detects out the colorectal cancer, how ai is helping humans to sway off the suicidal thoughts  and also how ai is been helpful to curb out in finding out the pre-diabetic patients in earlier stage .

But now i'd like to talk about how this ai is helping patients to find out cardiac arrest at the emergency conditions and also helpful in diagnosis . European Emergency Number Association ( EENA ) has recently announced their partnership with Danish AI startup Corti : a real time artificial intelligence assistant that has the capacity to rout out emergency conditions for cardiac arrest in Europe with the precise decision making skills with the intuitive training of deep learning mechanism concepts and adaptive neural network algorithm technique 

cardiac arrest emergency condition

Cardiac Arrest is a most common heart illness that happens when your heart suddenly stops pumping blood around the body and it is caused by abnormal heart rhythms called arrhythmia. It is not quiet an easy task for any doctors to conduct a check before it comes but some of the habits that causes cardiac arrest in high magnitude possibility. With the artificial intelligence that leverages the concept of deep machine learning with Neural Network Algorithm , now it has become quite possible to chuck out the pre-matured cardiac arrests so quickly .

Corti's main goal is to make sure that it detects properly and improve the output efficiency , accuracy of dispatchers diagnostic process with the deep learning analysis even the human brain could not able to match the accuracy of corti's out forte brilliance in recognizing results so accurately and quickly. The tests were conducted on 4000 emergency calls on cardiac arrests where it seems to be Human diagnostic dispatchers could able to trace out with 74% accuracy whereas Corti outsmarted those results with 94% accuracy which is stunning .

corti processing technique


How It Works ? 

 Working procedure of Corti seems to be more flexible and user friendly. When a patient or a by-stander or any tres passer makes the call . Corti will assist the dispatcher that listens to the conversation that triggers the signal for both verbal and non verbal communication such as tone and voice are easy to be considered to produce more effective result in approaching diagnostically with the digital assistance through deep learning technique by listening to sound stream to extract the most important and vital features which are necessary to process through reasonable frameworks 

When the tone and voice are recorded simultaneously , all the data which were analyzed or recorded during emergency calls are now taken into plot and automatically detected , analyzed by CORTI with this pre-defined trained meta data results to match the patterns and find out the immediate and effective results. This could be achieved only through Neural Network Training and Deep learning complexity which makes this possible through millions of database collected for research and comparing during analysis purpose will be now called for reference during analysis digitally 

As Corti is a smart digital assistant that has the in built and learns from the existing results more quickly and it predicts the critical situation of any patient based on situation , symptom descriptions and signal analyzing at the time of emergency more deeply via. audio and voice recognition. It then sends the alerts and notifications to dispatchers in handy through its intelligent reasoning in real time and predicting accurate and incredibly fast by combining machine learning models that utilize advanced feature extraction capabilities . It works on recurrent and convolution neural network approaches to examine a larger contextual input 




EENA is now in a verge to spot out the locations which are suitable for Corti to be implemented at the earlier stage than going wide and the director is very clear in his decision to extend its operations all over Europe later and then to the globe to show how ai is very successful in making decisions at the real-time scale during emergency conditions and it is pretty sure that this Corti is good enough to pick up the distress signs than humans more effectively is a challenging work and also risky too but good and smart initiative startup from the team to bring another limelight in medical field with the aide of AI is fiendishly brilliant & pretty insightful. 

SKETCHAR : this AI deep learning conceptual tutor directs the users to draw the pictures easily with help of AR & NN

Drawing is an art and doesn't come naturally to everyone , some people does it for profession and some for hobby. To teach drawing to others , one has to come with fundamentals of drawing to train the students to draw from the particular direction to get the flow well . 

It is not so easy for anyone to pick up this trait so easily not unless we are completely interested in it. When this has become so difficult task for anyone even if they interested to do, there is an app that helps the users to get to drawings more quickly and effectively through Augmented Reality , neural networks and Machine Learning (i.e.,) SKETCHAR app. 

This is just an application designed specifically through which the user sees the visual image on the surface are they planning to trace or sketch. Users will be holding the phone in one hand and sees the virtual image and the other hand is meant for drawing virtual lines which are sketched on paper which is been directed by app through augmented reality and neural network pattern recognition algorithm . You can draw with either plus signs or circle symbols to initiate.

sketchar app  and drawing pen

I was recently talking about Moleskin's smart gadget but now i'd like to say something about one app that really facilitates the drawing power of any user with just + signs as start. Interesting fact in this app is you can use it anywhere like you can draw it on notebook , paper , pillow cover , any where and you can get to learn easily 


You can just start up with plus sign and you will be cycled through point by point lessons that helps you to associate layers to draw . Actually ,it works very easily by converting the content which is rendered and the original is mapped and then we need to place it on the virtual object on the surface that requires more technical approach. When each layers are approached and separated , the process is been done through automated teaching of algorithm to camera to distinguish between everything 

sketch ar drawn content recognized
( image source : https://sketchar.tech )

As you can see from the above source that virtual image is located on the surface and its been snapped through smartphone which is been recognized and the paper is separated from background and its clearly separated from hands to surface and then its been detected and recognized of the drawn content. This app is an ai initiative and its fully powered with neural network training technique with the supportive of augmented reality. Not only that but also compares the false recognition and compares it out 

Whence this app knows how to separate from original objects / patterns clearly , the images are then latterly matched with original and the action of algorithm is been processed for image pattern analysis and it is very challenging app for any users to get in touch with. They also come with video tutorials for every users to follow up as per instructions shown which is very easy for anyone to understand.

Most interestingly , it is presently available for android , ios and also for Microsoft Hololens. For further more detailed information, please check it out https://sketchar.tech/ and for video tutorials of how to use instructions , please follow up this https://sketchar.tech/tutorials/, for Video understanding check out the video below 



Please subscribe to their official Youtube channel over here and get accessed for latest updates 


IBM's Watson powered voice assisted home's and car's will soon be in action

If you have watched Sherlock Holmes television series , you'd have noticed one prominent dialogue " hello my  dear Watson " to his buddy every time to close the case. Likewise, the tech giant IBM has named this Watson to do more assisted tasks through voice  as it is ai powered. Watson was under progress so far and it seems that IBM has officially launched Watson recently which has got more functions to do.Watson is not specifically designed to be brand conscious ( like Siri for Apple , Google Voice assistant only for Android phones , Bixby from Samsung ). This ai powered gizmo could be installed in any hotels , hospitals , cars and homes. Interestingly, when we talk of home assistance through voice, we could recognize Amazon Alexa  and Google's Nest from my previous posts. But this IBM Watson got our grip ahead into multi functionality assignments through voice assistance seems ti be quite interesting. 

features of ibm watson

location of ibm server

 ( Image source : Medcity News )

IBM Watson's partner Harmon demonstrates Watson assistant and this assisted service seems to be first done for voice assistant cyber security last year but now this function has been escalated to do wide variety of tasks that could be enabled through voice or text depending on the the device you use. Interesting isn't it ? , Whether its a home speaker or thermostat or  security systems or automated systems in car are all could be monitored and controlled through this multi faceted IBM Watson cloud powered facility through voice or text assistance. The most interesting feature of this feature is not only it will do actions based on the voice commanding assistance but also learns from the commands we assisted as it is also designed to learn from the actions and those preferences are remembered. As this IBM does our assistance tasks very effectively and also automatically by checking into hotel , rental car is coming on time , checking thermostat , your email checking also could be facilitated through cloud powdered IBM Watson to your phone automatically through IBM cloud based dashboard. It is like smart assistance enabled feature that you will receive notifications whenever the task is been performed but how ?. For that the devices should support  this IBM powered Watson and the account of IBM Watson could be enabled through IBM cloud account to get assisted and also data's could be accessed all over other Watson apps and you could imagine in future its gonna be fully powered in various gadgets as well. Please check out the video of how the IBM Watson works






ONNX : a combined team work initiative for ai developers to hover around framework easily

Human being's support are really mandatory for any field streams from planting , mining and on space as well. With the Machine Learning's day to day prudence in variety of arenas that humans consume time a lot , this deep learners does this task hand-fully in minutes span of time that are extremely brilliant . When companies like Microsoft, Facebook , Amazon join hands to frame work one need platform for ai researchers and enthusiasts ,it is truly a gift for the developers to build to put to practice. Months back , i was stating about Gluon , now i am going to state something more alike to that of is ONNX ( Open Neural Network Exchange ) a platform where re used trained network could be trained to use for multiple platforms . One of the foremost problem we face in this network while we develop is to choose the right framework. 

Though Data Researchers and Data scientists has large no. of options available , it is not so easy to port the trained model into proper framework. Even though it takes lots of tasks to train neural network signals to practice to retrain will be a stress making process for all developers to re-frame and upload it to proper channel. ONNX does this task more easily by allowing users to retrain and re-frame it through proper export without any problems that leads the users to develop and train them more quickly. Developers can import PYTORCH model in Microsoft Cognitive Tool kit or process any images in Tensor Flow flawlessly with this uni-platform. This inter operation enables the users to reuse of any models into any proper frameworks without pressure from scratch till end of the development.

onnx details in website

As you all know deep learning needs parametric training and it is not only lengthy , complex but also expensive one. In terms of time and complexity, its a round off process like if you pass each data set through ai, neural network is trained and evolved based on parameters such as input layer , hidden layer and output layer via. modern training process. If you know of neural network training method, each weight is associated with layers to be trained and processed for evaluation under correcting neurons through proper optimization say genetic algorithm or back propagation algorithm. For every evaluation , accuracy value is compared and manual adjustments are tuned manually or automatically.

ONNX inter operability training from pytorch to apache market

onnx converters and frameworks , runtimes

  ( Image Source : ONNX )

Each frame work development is meant for specific purpose and developers can perform any tools without compromising quality and performance and also based on ONNX community .Google didn't join yet but it is another landmark in developing the framework projection. It has supported frameworks and converters to do it easily , also you can import and export easily between frameworks without any hassles. To how to do the import and export task , please visit their official link https://onnx.ai/getting-started and to check out the end-end tutorials please check out the github repository https://github.com/onnx/tutorials. For more info. about ONNX news and other supporting stuffs, please check out the official website https://onnx.ai and for self understanding please check out  how Nvidia linked out with ONNX and they made it successful in the video below


Piccolo ai tech will now automate and also sense your every move through its gesture recognition technique

Artificial Intelligence is taking a forward linear advances towards modern's trends on commercial necessities. From the bot developments to spatial gizmos , the presence of ai has shown significant trendsetter even with the smart automation at home , the glimpse of technological bloom has magical presence these days with the convergence of research lab using camera and voice assistants around the house. Few weeks back , i mentioned about eye tracking device tobii , now the technology has gone one stop further towards motion assistance in home through smart facility

smart camera on wooden stand

This works on the principle of Vision assistance which works similar to voice assistant and works with camera and computer algorithms. Gesturing is one of the advanced tracking stuff which piccolo focused on . Gesturing through controlling things to turn on and off. Gesture tracking works more faster like we command on alexa. Company is very specific about making it lighter and quicker through motion sensing  susceptible to change according to the motions deducted.It has one more interesting function called autopilot which means it could able to adjust itself when owner is not there. Suppose if the owner went out or slept suddenly while watching , the control senses it automatically and dims the light and also controls the temperature and also volume raising / reducing  functions through enabled assistant.

human activity is sensed as gestures through smart camera

gesture recognition movement has been recorded

The Gesture Recognition Pipeline and Device Location Pipeline indicates not only the motion based technology but also indicates the devices on apps through deep learning and 3D rendering and reconstruction with the deep mechanism to identify people and apps at good speed. That makes the camera allows you to control the gesture stuffs and create a platform to build own apps and find traces of location . It is also controlled by app too. For detailed info. and product purchases, please visit their official website https://www.piccololabs.com/ and for more detailed information , please check out the video below




Vuzix blade glass is an AR smart glass that seems to be functioning more of Google glass rivalry

 Intelligence and technological advances has sparked up more momentary perception in terms of Augmented reality and artificial intelligence these days. Some companies are making it more severe to the public to make it more commercial sooner . Recently , i mentioned about tobii eye tracking gizmo , now i'd like phrase about some glass initiative like Google glass if you are not aware of has some wider features for public through interactive conversation in smarter way which we could not imagine. 
vuzix smart black glass with technical specs
Like mini smartphone , it has all technical specification like Android OS , Microphone , HD camera, touchpad , micro sd slot , usb connection and bluetooth functionality . The main core features of this so called alternative to google glass has visionary and technological pre-installation like patient data , direction mapping, information about weather , alerts and so on. This gizmo will be a perfect pitch to the geeks and it is a wonderful companion to smartphone users in many ways. It is such a hand free to the mobile users but it will be depicting the conversation with whomever we talk to . User doesn't needs to pick up calls to have conversation , with just a tap at the corner calls could be responded smoothly.


 lady touching side of the smart glass to tap for smart facility

smart glass that shows the person who calls us to our no. on screen

smart glass capturing photos of kids playing game on screen

This is just a similar model of Google Glass type with facilitating augmented type function but the only difference is it is compatible a with alexa amazon voice assistant with wave optics, which means its voice assisted too... it could be connected to headphones to control the function aside. It is also known from the recent news that amazon alexa is capable enough to communicate with other electronic gadgets say microwaves and electric vehicles , now the list will grow even more wider in future. It gives mobile accessory kit to users ; that could be easily accessed through bluetooth headphones and power to use them and wearables that could be assist through voice controlled system. Company promises that they does offer the product at cheaper price in future and also they explain that unprecedented access to data collection and information to retail supplies , manufacturing , logistics and medical and industrial purpose. It has a fascinating audio , video stream and playback facility with its magnificent specification but the price seems to be bit expensive as it costs 1000 usd right now .

vuzix smart glass types

In my perspective , Virtual Reality and Augmented Reality devices are growing faster and coming to commercial use for industrial and domestic purposes in future with lots of promising benefits to make us to have a feel that is close to reality in digital world . Very soon , this parallel illusion of such virtualism in digital world will come to our hands and dominate us and i am afraid will that be an end to our real world . Have a look at https://www.vuzix.com/  for products and purchases, to know more of what it can do check out the video below








Tobii's eye tracking is going to be next virtual experience future headset

 You might be aware of  augmented reality : direct or indirect view of  physical real world communication and it is keeping on track and also about Virtual reality through headsets to project real time environment , but if you take it bit further eye tracking could be a leap and this eye tracking facility , one could able to achieve highly integrated devices to more livable gizmo. This is a pioneer in VR technological concept with the fusion of sensory devices like tracking hardware & eye trackers. tobii eye tracker is a frontier advancing technology that allows computers to  understand what gamers are take off.  It is going to be integrated with next generation VR headsets.

tobii eye tracking technology

This technology seems to have its first integration at HTC VIVE at GDC and they showed up the benefits of eye tracking technology far more than Virtual Reality and the most special thing about these are it senses and reacts like mirage that reflects the avatar. If you blink , the eye tracker will blink, if you blink and also it tracks the sensory organs in your eye through controller and PC and does the tracking and motion actions with just a look.

laptop with eye tracker in service

Tobii team is very confident in making out the sensory based movements without using keys much is going to be future. When you just login to the computer , you no need to do anything , it just tracks your eye direction and even doing for game characters interactivity with favorite rendering. Eye tracking comes with wonderful pixel picture quality . Wherever you position your eye while you watch any video or play games or surf net, tobii delivers the best quality with more amazing experience. Sometimes , during gaming it is very harder for any gamer to look at the track , but if you have this tobii eye tracking service controlled through visionary perception and they did splendid work through powerful automation  in Virtual Reality experience tracking with wonderful feeling to next level like if you are real time player that fights with other persons and playing games in virtual reality concept.

tobii chipset
 ( Image courtesy : engadget.com )

  At present , more than 100 games are supported in this technology and it seems to be available for thinner laptops for few years and the company is making a new trend setter in gaming technology and it seems to be slimmer and it proves favorite rendering and it is like we are watching the future with better VR headset for further more info. please check out https://tobiigaming.com/ and for more video explanation , please check out the video below


To my personal opinion , development team has made a  benchmark in Virtual Reality concept like in four dimensional realisitic style

Robomart Bodegas is just another robot that delivers veggies to your home with just a tap

 Intelligence of computation of machines have become a testimonial benchmark to test bound their hype to next level these days. In the case of organic fruits and vegetables , we always need to rely on fresh markets or online stores usually. A California start up robotic maker Santa Clara's new product Robomart's main motto is to bring the fresh and organic products to the public to their door steps.

robomart bodogas car

  This so called robotic mart has its own specialty feature like there are racks in the car where the vegetables and fruits are stacked and arranged properly like in food mart stores. The most interesting part of this robotic stuff is it like tap and access facility.  Simply just tapping the button to access the closes robomart. When you approach near , the door opens automatically and you can choose the veggies and fruits you want and it will close it automatically , when the shopping is done. It is like Grab and go patent free technology will just send the receipt to the customers

robotic bodegas car specification and features

 This Robomart Bodegas is start of the edge technology that comes with autonomous wheels under autonomous car driving concept. and its also mobile friendly , electrical equipped with incredible environment friendly type and it is termed to be level 5 autonomous concept with HEVO wireless charging system with full refrigerated control which is totally programmed under Nividia AI concept comes up with full technical specification like Lidar, Cameras , CAN motion control system in full techno ai conceptual drive moving at 25 mph top speed upto 80 miles. These robomart bodegas has fleet management system that has ordering , routing , tracking , restocking and tele operation stuffs which could be done both automatically or manually. Overall , this ai initiative has a good stand towards putting it commercially to a good handy use. for further more deeper info, please visit their official website https://www.robomarts.com/. They are still waiting for DMV  license patenting from California and hope they will get it cleared soon. For video reference, please check out below




Teach codes to the kids with this interactive and fun Makeblock Codey Rocky Robot

 When i was in school , i supposed to study all programming languages to pass the computer science exams with the help of tutor and books . Those days , i was learning what is computer and how the computer operates through programming and what does that syntax means is all tho-roughed with the information stated in book and doubts cleared by tutor himself. These days , when technology has a leap towards robotic mechanics arena, robots are capable enough to do some mini stuffs and micro stuffs and also humanoid tasks as i stated already in my earlier posts. Now , i'd like to mention of one robot that does coding teaching to the kids more interestingly and also more productively like a toy learning 

codey rockey robot starting price in kickstarter

   The most interesting fact about this little toy like gadget interacts with the kids like a live cat stuff with some technical and hardware specification details . This is actually a kickstarter start up project aimed to hit consumers with the motto of teaching kids to develop codes by themselves through easy usability which means this little robot could help your kids to develop their cognizant and concept building logical skills for a code development to assign tasks to the robot by the kids itself without any help from any other programmers that is truly a promising project . It is a mini teacher that comes with GUI gives the children the instruction to develop and build by themselves . Codey rocky has a controller head with led lights in its tiny little head with A,B, C buttons that could be attached or detached easily like a docket

specification features of robot head part

   As stated in the specification details of this detachable kit , it is not only presented with sensors and led displays but also it has gyroscope , infrared receiver , gear knob and also speaker too. Interesting gadget that has a flexible facility that could be docketed and pinned wireless or via. bluetooth to our smartphone and tablet to perform certain tasks which are already configured inbuilt what it needs is a code to access and propogate it. It seems this tiny little gadget could perform plus functions as like in our smart phone.

programming ide platform of codey rocky robots ui

codey rocky robot attached to tablet sensing moisture

 codey rocky robot that control's tv channels remotely

kid holding robot  saying that he can design own games

       This handy portable GUI mini robot does many functions which are amazing and helpful to kids and adults as well. There are codes to customize for lights , tell weather report , see time and personal messages not only that but also with the help of this kit with bluetooth ( via.wireless ) and USB ( via.wired ) but also it does varieties of tasks could be performed with the help of this gizmo. You can
  • You can pin up with Makeblock Neuron and Lego bricks to generate ideas 
  • You can upload or download from any pc or smart device through pinning via. dongle 
  • It senses the light and moves towards the direction with the help of light sensor
  • You can write and assign the code to make it as musical toy with the help of buttons 
  • with LED display and internet connection enabled, it is capable enough to tell the weather report timely 
  • You can design own games through codes 
  • You can control tv remotely 
  • It has Colorful GU ten programming languages  that could be built and developed to assign the code and re arrange it anytime 
  • You could also assign the code to monitor actions 
infographic illustration of codey rocky robot features and benefits
With IOT principles and Artificial intelligence learning adaptability any kids and adults could easily perform a good handy coding operations themselves and also they can learn easily and more creatively. This affordable dynamic and creative embedding flexible mini paratroop sized robot does gaming and also worthy tasks which are promising to our kids in coming generation. It also comes with other accessories and smart functions as well. It is still under modular development and product initiative has been succeeded by team by bringing up this niche product to the public targeting 6 years age and above kids in the consumer market with clear vision of offering them a mental ability to think when they grow up with the GUI functions and controls as toy and also robotic tutor as well. Please check out the project in phase documentation and other details over here and if you want to know how it operates and more details visually , please check out the videos below


    The kickstarter project is gonna be a fresh trendsetter in the robotic / ai in gaming /coding / home use commercials sooner with more benefits and features that could give others a touch to start something more handy robots that could be used in home and in office for other tasks as well. Worth the shot for this 49$ start up kit . If you are looking for further more accessories , please look over the cart section before purchase over here https://makeblockshop.eu/products/makeblock-codey-rocky and also in the above link as well. It comes with user manual and also interactive guide to customize and play random games. To support and back up this project Please visit this website  to encourage such a wonderful project



Google AI and NASA tied up to discover the other exo planets

 NASA is formidable in taking a geo-spatial and exoplanet discoveries in every astronomic moves with their incomparable standards to the public . With the team of scientists and researchers working towards one communion to achieve their mission targets on stand for past , present and future. The tech giant Google is on hands with NASA team to integrate to unity to make it even quicker and more possible ways. When we think of AI these days , the computational probability is sharpening its traits in the field of astronomy too. If you are totally aware of Kepler 90i - exoplanet discovered by NASA is discovered by Kepler system . Now, Google AI is  taking steps even further to head direct the move to find out the other exo planets in Kepler mini solar system.

kepler 90 solar system in comparison with earths inner solar system with  black bacground showing earth , venus , mercury planetary location

( Image courtesy : Science news )

This hot and rocky planet was discovered by Kepler space observatory telescope with the help of Google's neural network algorithm  and light readings were captured very interestingly. Scientists believed from the observatory that Kepler 90i does have solar system like ours . The discovery came to limelight with the help of Ai and that Kepler is seemed to be a sun like star which is just 2545 light years from earth and it orbit its star once in 14.4 days. Neural network algorithms are efficient enough to locate the weakest signals and the data's will be collected in database for collection of repositories. Kepler's system is like smaller version of our solar system and for past four years , the team seems to be collected nearly 35,000  signals. 

Here Google AI's help is rendered with the proper training of data's as neurons with at least 15,000 plus signals as samples and that is been iterated with the value adjustment . With the human presence , automated tests been taken with the neural network back propagation algorithm training till the errors are minimized . As Kepler's signals are been collected by system , it is been tested and verified iteration-ally with so many tests to prove the authentic confirmation and seems to come out with 96% accuracy. 

neural network with input layers and output layers in circles and arrows
Earlier, Neural network were just trained to search inside the database and never been used for further developments. Recently, google team promised the team that they will come up with better ideological concepts that could help them out to curb out , which could not be easily identified by our eyes or with telescope alone. Kepler 90i is a promising one to the team and they also state this system just has more similarities to our solar system  and thus helped them to find out sixth planet in Kepler 80 orbiting system and is comparatively smaller .

   Senior Software engineer Shallue is in joint hands with Google AI team came up with an idea to establish the connection of data with neuron network to get trained and it paid well to the team . According to the research paper published in The Astronomical Journal, Kepler was in deep exploration of signal adjustments and the results shows the values of Kepler mission are deeply concentrated.  Google AI seems to be doing its task very well by exploring other exoplanets by teaming with NASA to achieve great results for future exploration. Though neural network has to be trained a lot to understand the signal patterns more accurately , it does its task very minutely. They also made Reddit AMA section recently.  For better understanding, please check out the video below 






This small mechatronic pongbot just owes us a short time fun for boozers

   Do you drink beer ? and do you love to have some fun times with the cups and bottles like kids ? Here is a small robotic kit that actually does the holder that aims the moving target. If you are aware of ping pong , the ball just moves in one direction through hits , but this gizmo does the spin , darts and randomly changes the direction but also senses perfectly by not falling under the table.

red colour ping cups with pingbot

This edge sensing technological kit moves in all directions by holding a single cup upto maximum five cups on the holder above the kit with holder caddie. You just need to stand apart and throw the ping ball on to the cup while its moving . Most interesting thing with this is those cups will be filled with beers and the opponent will be throwing the ball on your cups . You can either control the pongbot via. remote or leave it to automatically . whilst it moves , the user will be throwing ping ball on the cups and if the user throws successfully , the opponent will be taking that drink and this is not an hard task. This is very funny game and i could remember my old days playing these kind of games by adding sand in the cup by filling with balls through hits .

black colored pong robot with holders


This kit has LED lighting ( Red / Green ) and the player doesn't even know which direction its moving , it has edge sensor technology . You can play this game during any parties or when having fun time this works under motorized remote sensor control mechanism which has batteries and cups coming with the kit for free of charge and also its very easy to operate . You can take it anywhere to the party and play it with no additional stuffs to be installed. It is easy to use and hold and hassle free with the power switch

black color motorized pongbot with holder and remote


Actually to play this , there is no specific rule or instructions to play this game. As long as you have this pongbot , you will feel a good time killer in playful way. This is a cute 40$ kickstarter project that is absolutely a gaming one which is quiet funny with limited options. This is a pure mechatronics project works under pan mechanism . It doesn't have any regular pattern , but it is very tossy. Fore more information about Pongbot , please check out the video







First virtual politician chatbot responds promptly via messenger.

      Artificial intelligence is no longer a hypothetical topic to anyone these days, it seems the rise of computational intelligence can power anything in mere future even in the filed of politics and governance. Computational neural network is having an extraordinary ability to develop and train the platform to analyze and decide to take further steps in logical and conceptual decision making skills. The smart deep learning ai politician  named " SAM " does this political task

sam virtual politician logo with yellow background

(Source : http://politiciansam.nz/index.html)

      This is smart virtual assistant which does the political disguise and communicates with us in more stylish and friendly way like customer service team 

chatbot responding the user with responses
sam chatbot responding the user with responses


   This virtual assistant is having a complete political information and does like automatic bot task but all information related to Newzealand constitution through facebook messenger . For commanding the input just check out  www.politiciansam.nz/helpchat.html for the list of questions. You can ask about politics , education , childcare , healthcare , business and et al. Like google assistant , this chatbot does the task so promptly but only with the limited knowledge offering the response through collective info. It seems that machine learning is taking traits of our individuals slowly and sequentially in every corners at the initial stage by putting human decision makers to concern of ai bit more a healthy and advanced competitor even in election campaigns in future it seems. 

features of sam ai

This is just available only for Newzealand citizens to know about the political and statistical facts but any user from anywhere can just try asking certain questions which are mentioned in help page of the website.SAM ai will be expected to participate in 2020 election it seems says Gerritsen . For more information, please visit politiciansam.nz 

Gluon : AWS & Microsoft in hands to introduce Deep learning combo for developers

 It is no wonder to me the two tech giants like Amazon and Microsoft are hand to hand for first time ever to develop a much better machine learning accessible for developers . This must be an open source interface that concerns on deep learning into more deeper interest . Gluon is termed to be a built unanimously to train neural networks in terms of data, model and algorithm to make sure that is easily accessible to any developers to create and research more and its like a digital library platform for the developers


gluon logo with api specification

 ( Source : github.com/gluon-api)

Gluon is an easy-to-use and build interface that is open to developers and researchers in machine learning and deep learning conceptual training . As the purpose signifies it could be built in with any structure for optimization and layer definition . As this algorithm trains the model to easily understand the data patterns and automizes it accordingly. Why its very significant to the developers ? Because one good reason its very easy to understand and implement like installing a plugins in our browsers , secondary reason is it helps the developers to interact , build and develop to make a quick prototype. It seems that developers no need to shy away in making up the projects from the api which is openly posted to general public on github and there instructions . Microsoft got this connected because they are already in the verge to make numerous projects from ai to boost the productivity in various fields in future . So they come to hands in join with Amazon web services to make it possible to compete against their competitors .

Gluon can be used for training the neural network to train and resize according the users convenient wants. It compasses the both neural network and training model and training could be performed so easily with no hassle free to develop any app to interface it and to deep learn it, training is a predominant factor and the researchers are saying that it has become a smart trainer to students and developers to prototype the model so easily. Recently , Docomo has made a step further towards showing an interest about prototyping via. Gluon is a linear platform for ai research developers and data analysts to be honest more. Newyork times recently reported that the two tech giants shaking hands at their back of sleeves to form an artificial cognizance research development to carve the future to combine such moves has become mandatory to sprout the redundancy to strike out rivalries by waving hands at Cortana from Alexa through intelligent approach from both the teams . Now Gluon is technically occasioned with Apache Mxnet to heed up and the tech team from amazon are about to make a further deep movie with other companies through partnership for ai research and development collaboration. For further more info. Please make a check out here

Google's learning software is now prototyping itself to train

Not in one day i could not see stop seeing about google and their day-day news is really marking beyond the range we expect . I was talking of Google Assistant yesterday how does it tracks out the unknown songs or tracks that is running at the back via what song is this command line, but now i am about to mention about the google 's machine learning program had made a self awareness that replicates itself . Couldn't able to get any idea what i am talking of ?. Google always amazes us with fresh products and announcements that really seems like we are already living in future today. In May, Google revealed an automated project called " AutoML " from the name implicates that this project is specially designed to develop other Machine learners by itself . It seems this project has already shown the tease by beating up the human intelligence in building up a game machine is what google research team is saying of. In this one , google has successfully implemented from image recognition to machine learning so far and it is very significant for everyone to be aware of. 


neural training itself to create more neurons in flowchart

the controller is training and verifies the model in ann

( Source : Google Resource )

   Above here is a flowchart of symmetry that happens with AutoML program that encompasses the structure of training the network that is associated with it. To get more accuracy in deep learning theory , this AutoML creates data through layers of various neural networks. From the above flowchart of Google net you could able to visualize that the Child model architecture is first proposed and will be assigned for specific task then latterly trained , controlled and feedback will be send to the other architectures after thousand times of training that could happen in fraction of minutes . Later on , thie above training and feedback process is repeated to other layers and does this learning task symmetrically more better.  It seems that AutoML project is outsmarting human's intelligence in various tests with better accurate results that leads this team to continue this project furthermore. 

   To accomplish this , the team is working very hard in implementing all recognizing algorithms among-st with reinforcement algorithm . The team says training part is just  a trail, the main headneck task is to mimic the human brain like stuff which needs more training and tuning layers into proper adjustments with specific inputs. So far this AutoML project is been implemented for image recognition and speech recognition tasks which we already knew of it. Now the deep learning technique is even going further to put a web of strings in terms of patterns and layers as neurons to upsize it like a kid to grow up . The process called "meta learning " or learning to learn to speed up the networks to perform better architecture. This task does looks simple but not easy , team were pretty complaining about the mundane of repetitiveness to check which layer or which one will perform better through no. of training and iterations 

   Google CEO Sundar Pichai is very much confident in this project and he believes that this project will be further implemented and executed for variety of learning algorithms purposes to put to use to humans in various numerous ways in future. Let us hope for the best , as we can see now what google is capable of more than we thought and they gonna fuse our lightbulb with the help of transcending artificial intelligence infront of us to shake hands in few years time is gonna question our roles in future. For more understanding about the concept here, please watch out the video below :)






AI can fool captcha ?

 Artificial intelligence a.k.a machine learning has become a niche to our day to days life. Decade back we used to authorize websites with user authentication details , where password is really mandatory for every person to access his / her account . When so many spamsters tried to knock it away , there comes two way authentication and captcha (Completely Automated Public Turing Test to tell Computers and Humans Apart ). Many of them might away why captcha is been placed at the bottom of user registration or sometimes at the login page. Many people might be aware of what captcha is , for those who are not aware of it , i'd say that is a security authentication that distinguishes human from machine learning through sight and type proofs 

captcha with word verification dialog box

As you can its a automatic program that shows jumbled alphabets , numerals and sometimes pictures to authenticate user and to distinguish from getting spammed or botted. Likewise, AI is making a progress like human to prove its authentication by outsmarting this activity . Computer scientists and researchers recently developed an algorithm that mimcs and visualizes like human brain that improvises visuals apart is something more humanish is what i could say. Normally captcha is the visualizing stuff'and computers are not easy to pass this test but somehow this google even pointed out that solving the recaptcha test is complicated to even human beings , but computer researchers and scientists from the  Vicarious proclaimed that they could able to tuck it away using special technique 

Normally, Neural network techniques , image processing technique and recognition algorithms are widely used in most machine learning concepts , but this team has made it further to develop an algorithm that creates Recursive Cortical Network a kind of pattern technique to mimic the actual visualizing process as like in human brain more better than neural network , as neural network needs literally thousands of images to be present and processed and correct errors takes lot of time to tune the accuracy but this seems to be working well and even requires less computing work than neural networking , which is a vantage point to the team . 

A Captcha results table

( Source : Vicarious )

Human brain has the ability to visualize the images and objects , once they see over and patternize it , likewise this RCN is capable enough to make the patterns and visualize itself and even have the ability to analyze the pixels of an image matching the outline . In 2013, the same team has made an announcement that they cracked the text captcha result with good accuracies which are tested by google , yahoo , paypal etc., Now it has gone even more wider to prove its recognition with more analysis to improve its efficiency . It seems that Ai has taken deeper leap into crack down these security authenticators to by pass human identities very sooner . To understand much more better , please check out the video below



Google's Deepmind AlphaGO Zero is really smart enough to beat the world class Go player without any support

Artificial intelligence is shining its sheer intelligence in many modern ways these days even more smarter in beating up the human intellectual proficiency and brilliance in terms of logical thinking , accuracy. Yesterday i was talking about Ai in fooling out the CAPTCHA test very effectively better than humans , but today i'm going to talk about a deep mind learner machine that beats out the human in this famous chinese game GO. This game has been played for more than 2000 years and its very famous in Chinese and all over the world even in pc

Go play board with black and white coins

If you have played this game , you might be knowing how to play , it is more of tic tac toe like type. Okay let me get straight to the point , Deep mind has been playing this game since 2016 March in China with top players and it amused the audience by beating the players in  four of five rows which is not easy for any human being to beat another human who is equally or more smart in playing particular game and it is also been reported that nearly sixty million people went to watch that event and its been recorded and it has given pride to its developers around the world with its intelligence and creativity and its also been said that its not beaten the opponent using brute force method also it has diligence to beat up with creative decisions is what the makers say of 

Google's Alphago with its logo

Atleast in my earlier posts , i was telling of artifical intelligence's training methods to learn and adapt using neural network technique  but this one never been trained to play that game. It just trains itself like i stated about RCN two days back. It is a sheer move and fresh experience to AlphaGo Zero through self adaptive learning technique like RCN . Although its never been learned , but it was playing random and system used novel based reinforcement to encompass the technique to combine neural network with powerful search algorithm to each time when it plays. Modern games like Atari been tried by this and this GO game is a challenge to artificial intelligence as its only played by experts very efficiently which impressed the players and other experts . Alpha GO Zero beaten its predecessor AlphaGO in all 100 games and its quiet a powerful performance in the break through of AI field to bring good impressive results to team . One important point is its been not trained and this deep mind successor learned ancient moves in this game by independent. It is not so easy for human's tendency to learn every moves so quickly and adapt to it so fastly without even trying . It is also been reported by the team that AlphaGO Zero learned and played itself nearly 5 mil times before getting into its predecessor . Could we even imagine with that amount of times ? Impossible in practical situations.

Co-founder and CEO of deepmind program Demis Hassabis says that he is very impressed and so happy with the way the deepmind's AlphaGo team is been setting upon  and he finally settled down with these words saying to the press that " Ultimately , we want to harness the algorithmic breakthroughs like these to solve all real world human problems like protein folding or designing new materials . If we can make this same progress on this problems that we have with AlphaGo, it is potential to drive forward to human understanding and positive impact on our lives

Go players in chinese tournament with google alphago zero ai

( Image courtesy : Science news.com )

It is interesting and even amazing to get to know that machine learning from Google's DeepMind AlphaGo Zero has mastered the game so effectively and even learnt ancient moves very fastly on its own through sheer learning technique . It is a next level improvement from the team to put that deepm ind to use in such effective parameter level and tune better than its predecessors . If you want to know of what is about deepmind Alphago Zero is , please watch out below video as what the team has to say to us and also please watch out how it has beaten the world GO master in the tournament and if you would like to get to know of more of deep mind program , please visit their official site  and please take a look at the official blog for further more detailed info. about AlphaGo Zero over here