Author Archives: Mayra Luna

Machine Learning In Our Everyday Life

Related image

When we think about artificial intelligence we might imagine something out of a science fiction movie. However, we might not realize that in one way or another we are using some type of machine learning in our day to day routine. Machine learning is a form of artificial intelligence which allows computers to learn from examples rather than having to follow step-by-step instructions. We encounter machine learning systems daily through our smartphones and our computers, and whether we realize it or not we have become dependent. One thing that’s for sure is that machine learning is already part of our everyday life.

The following are just a few examples on how we constantly interact with ML:

  1. Commuting: Through something so simple as using our GPS or taking advantage of online transportation services such as Uber or Lyft. Using location data from our smartphone, Google Maps can analyze the speed of movement or traffic at any given time. While we are driving, our current locations and velocity are being stored in a central server for traffic management. Our data is then used to build a map of the current traffic, thus giving us route suggestions and estimated time of arrival. As for App based transportation services such as Uber, machine learning systems are used to determine an estimate price of your ride, to minimally optimize your wait time, and to optimally match you to the best service.
  2. Virtual personal assistants/Voice to text: A standard features of smartphones today is voice-to-text. Using spoken commands to ask your phone to carry out a search, or make a call, relies on technology supported by machine learning. Virtual personal assistants (VPA) such as Siri, Alexa, or Google Assistant are able to follow instructions due to voice recognition. Machine learning is an important part of VPA since they collect and refine information, and later that data is used to render better results in accordance to our preferences.
  3. Social Media: From personalizing our news feed to ad targeting, social media platforms use machine learning in many ways. Facebook uses AI for facial recognition, so when you upload a photo, faces are automatically highlighted with suggested friends to tag. Instagram, which FB acquired in 2012 uses machine learning to contextual the meaning of emojis. As for Snapchat, the facial features (lenses) track facial movement allowing us to use animated effects or masks that adjust to the movement of our faces
  4. Online Shopping: Think about how Amazon suggests certain products as “customers who viewed this item also viewed”, machine learning is the technology that helps deliver these suggestions through recommender systems. By analyzing data about what customers have bought before, these systems can pick up on patterns in purchasing history (On the basis of our behavior, items liked, items added to our cart, etc).
  5. Music & TV Streaming Services: Recommender systems are also used to suggest movies or TV shows on streaming services such as Netflix. These systems use machine learning to analyze viewing habits and on how we rate that show/film. Music streaming services (Spotify/Pandora Radio) also use machine learning to suggest music.

 

Ethical Challenges in Machine Learning: A look at Advertising and Ethics

Image result for machine learning and ethics

There is no doubt that Machine Learning systems are transforming our lives, creating both a positive and negative impact. Companies such as Amazon, Facebook, IBM and Microsoft have taken advantage of the boundless opportunities artificial intelligence has to offer. However, businesses and governments are relying more and more on ML systems to make important decisions, raising questions about fairness, ethics and morality.

Despite the “optimism” that surrounds AI, we must be aware of its potential outcomes and the ethical challenges. Increasingly, technologists and policy makers are dealing with these questions, asking themselves how can AI and individuals ethically interact.

Some of the issues we may face include:

  • Unemployment: Labor is concerned primarily with automation. As we continue to implement machine automation, certain jobs (predictable physical work) could disappear. While this might sound like something bad, we must also keep in mind that new jobs will be created due to this disruption. The real question is how are we going to educate people for these new jobs?
  • Inequality: By using artificial intelligence, companies can drastically reduce the reliance on human workforce, which means people will work less hours (therefore make less income, broadening the labor gap).
  • Humanity: Humans are limited, while machines have unlimited resources. How will machine learning affect the way we behave and interact with others? Will we soon be interacting with machines as if they were humans? Will machines adapt to cultural norms?
  • Algorithmic Bias: Blind spots or biases in the algorithms could lead to discrimination against certain types of people (As seen in AI systems that can tell if you are homosexual).

So what precautions are being taken to stay on top of these issues? Recently DeepMind created a special research unit devoted to investigating ethical issues surrounding Artificial Intelligence. The people at DeepMind believe that these systems should remain under human control and be used for socially beneficial purposes. Elon Musk launched the “OpenAI” Institute, which is a non-profit research organization that “aims to promote and develop friendly AI” in such ways that will benefit humanity as a whole.

Also, a partnership on Artificial Intelligence has been formed by Amazon, Facebook, IBM, Google, and Microsoft (Apple joined at the beginning of this year). As best described in their web page, this partnership was established to advance the public’s understanding of AI, as well as formulate the best ways to implement this technology.

Image result for partnership on ai

 

Ethical Impact on Advertising

Artificial intelligence isn’t inherently good or bad, it is in the way we use it that determines its nature. Machine learning algorithms come into use across multiple fields, such as advertising. AI systems have the ability to collect and analyze data to determine human behavior. Every time you click on a link, google search something, every like you make on Instagram or Facebook, your location data from your cell phone, even your credit card transactions can give huge insights into a persons habits.

The advertising industry is obsessed with understanding human behavior, therefor taking advantage of these machine learning systems. However, as discussed in this article, AI raises many questions in terms of ethics. Imagine advertisers knowing us better than we know ourselves? The more they can understand us as individuals, the easier it will be to persuade us.

http://http://www.slate.com/articles/technology/future_tense/2012/01/behaviorally_targeted_ads_and_the_ethical_dilemmas_behind_building_consumers_into_ads_.html

The article above mentions the ethical implications of adding consumers into commercial advertising. In other words, through facial recognition and other technologies, behavioral marketers will be able to access your Facebook pictures, create a 3d image of you and insert it into ads. This obviously raises various questions on privacy, consent and data security. Imagine walking by a billboard with an imagine of a person on the beach. Now imagine that the person on the billboard consequently takes your physical characteristics. Creepy? Yes!! However, as with every other new disruptive technology, consumers will eventually learn to adapt to these marketing techniques.

 

 

Machine Learning and its Implementation in the field of Education

Machine Learning: How important is it for Education?

When it comes to new technologies and resources there are always questions and doubts. Most of the time we focus on the negative, without realizing how many benefits it can bring to our everyday lives. When it comes to Machine Learning and its implementation on education, many may wonder: Will teachers be replaced by robots? Not likely… Instead this form of artificial intelligence will be able to positively impact and enhance the teaching/learning experience.

So just how exactly can machine learning help in education? Well for starters lets take a look at a teachers workload. Teachers prepare class materials, grade homework and tests, provide feedback to students and parent guardians, etc. Machine learning systems can give teachers more free time to actually teach and mentor on a more personalized level, instead of focusing on the never-ending grading and lesson planning. In addition, the current learning system consists of a “one size fits all” model. We are in desperate need of a system where the educational experience can be personalized to each students abilities and needs. Machine learning could address these issues by collecting and analyzing the data generated by students, identifying meaningful patterns and transforming that information into structural knowledge. In other words, when the student interacts with a digital learning platform, the machine learning system can accurately predict and better assess that persons educational level, being able to tailor specific material for that individual.

 

IBM’s Vice President of education innovation for Watson recently shared IBM’s vision of smart classrooms, which are cloud based learning systems that can help teachers better acknowledge each students strengths and weaknesses, being able to determine what type of content to give that student and how to overcome their learning challenges.

Another one of IBM’s Watson projects is with Blackboard Inc. (announced last year), where they are partnering up to develop innovative educational solutions.

Other Uses of Machine Learning in Education 

This article goes over specific roles and examples of Machine Learning and its use in education. OVERVIEW 

  • Content Analytics: machine learning platform that optimize content modules. (IBM Watson, Gooru)
  • Learning Analytics: Focused on tracking student knowledge and enhance their learning environment. (ALEKS, Dreambox, Reasoning Mind, Knewton)
  • Dynamic Scheduling Platforms: Follows the learning patterns of the student and schedules him/her with an appropriate teacher. (New Classroom)
  • Grading Systems: Systems where students learning and knowledge can be scored (automatic grading). Ex: Turnitin, Writetolearn.
  • Process Intelligence: Focused on analyzing a large amount of data and identifying new opportunities. (Bright Bytes Clarity, Odoo, Jenzabar)
  • Predictive Analysis: Helps data mining jobs. Used to improve retention, learning and application. (Edulog)

Machine Learning: How it affects the Medical Field

Image result for machine learning and healthcare

Machine learning is all about data recollection and data processing. Many industries will be affected by this branch of artificial intelligence, the medical field being no exemption. Its capability of analyzing huge amounts of data will eventually replace a great number of human jobs, even those that require a higher education. However, as explained by Medtronic CEO Omar Ishrak, the real value of artificial intelligence is in making more efficient use of human resources in healthcare.  For example, IBM’s Watson can read 40 million documents in 15 seconds, optimizing time and performance. Healthcare has loads of data: Test results, consultation notes, scans, appointment follow ups, etc., creating an ideal environment for the use of Machine Learning (More data, better results).

Other application of ML in the medical field is the ability to personalize treatment and help in decision making. Insightin Health has recently launched an AI powered app where patients are empowered to make their own healthcare decisions. The In360 app incorporates user behavioral trends as well as real time data collection. Also, Machine learning is being used as an imaging and predictive analytics tool. Through the combination of Azure Computer Vision and machine learning, InterKnowlogy was able to come up with a solution for the early detection of Posterior Urethral Valve Disorder (PUV).

Machine Learning and Healthcare