Home / Articles  / Machine Learning: A Love Poem by Our TSA Experts, Part III

What is Machine Learning and How Does It Affect Search Marketing?

Computers have traditionally used a fixed set of instructions, called a program, to calculate inputs into meaningful outputs. Until recently the programs were written almost exclusively by humans. People knew what inputs and output they wanted and would manually teach the machine to perform each step.

Machine Learning, often referred to as artificial intelligence (AI), broadly describes computers and machinery that can program or reprogram themselves over time, with minimal human interference. Although this task has always been possible technically, it has only recently become a reasonably feasible undertaking, thanks mostly to advancements in computing power and efficiency.

Revelations from Google I/O 2016

The machine learning insights that sessions Google announced at the Google I/O 2016 yearly developer sessions were numerous.

Machine Learning: Google’s Vision

The first session is lead by Tom Simon: San Francisco Bureau Chief, MIT Technical Review. Tom interviewed three Aparna Chennapragada, Product Executive at Google, Jeff Dean, Google Senior Fellow in the Research Group, and John Giannandrea, Google Senior Vice President.

Some highlights from this discussion:

Areas where Machine Learning can grow:

  1. Learning from smaller numbers of examples
  2. Transferable knowledge from one specialty to another

Where is the significant growth?

  1. Language understanding
  2. Translation
  3. Conversational learning

Areas where Machine Learning can grow?

  1. Unsupervised learning
  2. Learning model learning (learning system to build itself in the neuronet.)

Where did the machine learning advancement come from?

I suspect the major change will derive from a few key insights rather than an extensive system of breakthroughs.

A Building block

A significant advancement would be Hand-eye visualization.

Machine Learning Spring

Until a computer can read, we cannot actually declare that it is Machine learning Summer. This translates to mean that the computer can read the text and then have the ability to paraphrase the content in the same language.

Humans are suspicious of Robots

There is a lot of talk in the press about the mistrust around AI. How do you build trust between machine learning and the public?

The conclusion from the session group is that machine learning is decades away from anything resembling AI from pop culture: Stanley Kubrick’s artificial intelligent robot Hal9000 from his picture 2001. Or the main character Michael from the Stephen Spielberg movie A.I. or even the droids from Star Wars. The key to helping people gain comfort is by researchers exposing the public to where the technology is now and not to dismiss concerns, but rather allow them to see Machine Learning in action.

Additional Google I/O 2016 Sessions

Breakthroughs in Machine Learning – Google I/O 2016

In this session each of the session speakers address an area where machine learning has allowed research engineers to advance past a particular technical roadblock. The talk centered around the struggles with Voice search and how to deliver better results to all humans in every language.

Machine learning & art – Google I/O 2016

This session focus was on how machine learning can become a great assistant for a code artist. The session speaker demonstrated an interface that allowed him to explore thousand of world art pieces and generate from their parts new artworks.

Machine learning is not the future – Google I/O 2016

The final session lead by Rajat Monga, Engineering Director at Google confronting the notion that Google believes that they might be the company that will be “AI First” advancing their current mantra “mobile first.” Rajat discussion where Machine Learning is the Machine Learning Spring and what will be required to develop the technology to push ML into the Summer of ML research. Rajat talk postulates that Machine Learning in the present will define Google’s success or failure.

Bradley’s Favorite Articles on Machine Learning:


Dirk’s Favorite Articles on Machine Learning:

Post a Comment