Working Up

by Dwayne Phillips

Ask for information. Use it to change me—not the other person.

I teach classes on a few topics at work. The best thing I can do afterwards (a few days later) is ask one of the students questions: “What is this? What is that? How would you do this? What is the definition of that?”

(1) Listen.

(2) Act.

Both (1) and (2) are essential. Let’s focus on (2) in this post.

When the answer is incorrect (there are, unfortunately correct and incorrect answers to factual questions), how I respond is critical.

Two responses from me (there are many possible ones) are:

(a) What? Weren’t you paying attention? How could you miss that?

(b) What? What was I saying? How was I teaching? Why didn’t I communicate better?

The difference between (a) and (b) is the difference between “you” and “I.” I can change myself. I probably cannot change the other person. Refusing to admit such is my problem, and I should get to work on me.

The same goes for meetings, speeches, essays, blog posts, and just about everything else. I can blame the recipient or I can work on me. And I can work on me without blaming me, but acknowledging that…

To write is to be misunderstood. (The same holds for teaching, speaking, etc.)

by Dwayne Phillips

Jeff Bezos and Amazon have received much praise for their use of written essays at meetings instead of PowerPoint and bullet points. Great, but not new.

This is a good piece about Amazon’s use of written memos at meetings. Keep the thinking focused. Don’t waste time. Think. Communicate.

Excellent. Great. I love it.

That is why I used it back in 2003 on a government project.

I doubt I was the first. I know Amazon wasn’t the first, because I beat Jeff Bezos to it.

I didn’t become rich and famous, and neither did anyone else on the project. We did, however, save the taxpayers a few million dolalrs, and that was a pretty good thing.

by Dwayne Phillips

It is just a “figure of speech,” but perhaps we would understand better if we used the correct words in the field of artificial intelligence.

Here is yet another article about “building” a neural network. It is a good article that explain neural networks well. It shows some sample source code to do so. I like it.

All except this concept of “building.” This does not show how to “build” a neural network. It shows how to simulate one in software.

I would love to see someone build a neural network. That requires hardware (much of it analog, not digital). An actual neural network built by someone would be a joy to see (at least for me).

Until that joyous day, we will simulate them in software. I know, it’s just a figure of speech, but perhaps, just perhaps…

by Dwayne Phillips

Somewhere, somehow we forgot what a good approximation was.

We used to tell this “joke” a long time ago in a place far, far away.

Consider the situation with an amorous young gentlemen on one side of a room and an amorous young lady on the other side of the room. Every two minutes these two persons reduce the difference between them by half. At no point in time will they ever touch one another.

The above has to do with the theory of limits in mathematics. It is true. The two persons will never touch given what is described.

The punchline of the joke is: Yes, but after a few iterations they are close enough for a good approximation. (No one ever claimed the joke was funny.)

Now back to reality…

Some folks at MIT discover that most neural networks have 10 or 100 times more neurons and layers than necessary. The persons building the neural networks have forgotten or have never heard about the good approximation. This is partly due to the near-zero cost of processors and other computing machinery. I believe it is mostly due to the lack of humor in higher mathematics courses taught in colleges today (we do still teach such courses, don’t we?).

Good approximations are often good enough. Especially when using data that isn’t that good. Consider the input quality and the needed and achievable output quality.

by Dwayne Phillips

We often use words that were used in a different time. If we understand that use, we can better understand what we are trying to do now, at this time.

I overheard some ladies talking about arranging meals for groups of persons at charity fund raisers. “If someone is really hungry that day, then we can…”

“Really hungry” at a fund raiser? Persons who come to fund raisers have enough money to give large chunks of it to other persons. Rich persons hungry? Really?

“Really hungry” is a phrase from another time when persons were actually, really hungry. They had not eaten in four or five days.

“Really tired” is another such phrase from another time when persons were tired from hunger and dehydration and yet had walked 20 miles a day for a few weeks.

We use these and other words when describing persons who sit in front of PCs every day in air conditioning while wearing new, clean clothes and having plenty of potable water on hand as well as micro-waveable food and…

“Those are just expressions, figures of speech. We all know what we are talking about.”

Maybe. Maybe not.

Sometimes—even today—we build systems for persons who are actually really hungry and really tired. Sometimes we build systems for the typical office worker in all those aforementioned nice conditions.

Let’s use different words for different situations. Words do have meaning.

by Dwayne Phillips

I know what I’m doing. You don’t. I can have fun. You can’t.

As a kid, I used to see wild stunts on television. They always prefaced them with, “Don’t try this at home.” Why not? Why could they do things, but I couldn’t?

Of course they had years of this and that and expertise and equipment that I didn’t. It seemed, however, that they were having all the fun. They were experimenting. They were learning. They were growing. I was sitting on the floor staring up at a black and white TV doing nothing, going no where.

“You need to learn a bunch of things before you learn this.” I think that is what they were telling me. But how would I learn if I didn’t try what they did on TV?

“But you need to go slowly. This could hurt.” As if going to first grade at age 5 with a bunch of kids who were all bigger than me and none of whom I had ever met before wasn’t going to hurt—a whole lot.

Everyone else needs to proceed with caution. Everyone else needs to go slow. Everyone else needs to watch me for a long while before…

Really? Who decided that?

by Dwayne Phillips

Really smart persons continue to blunder. How?

I read this article recently about how women are hurt more often and more severely than men in automobile accidents. Why does this happen?

The crash dummies used in safety tests work well, but are almost always the size and shape of a man. Crash dummies the size and shape of women are rarely used. More than half the persons riding in automobiles are females, yet we predominantly test for males.

How can we be this stupid?

Facial recognition systems don’t recognize black faces as well as white faces. This is because we don’t use many non-white faces in training these systems. The percentage of non-white faces in America grows and grows, but we don’t use these.

How can we be this stupid?

I could go on with more examples from recent and distant history. We blunder and botch these things that anyone who doesn’t know anything about the specialties can understand as stupidity. The experts simply don’t see the most basic thing.

It is a great idea to have an outsider, someone who doesn’t know anything about what we are doing, walk in the door, look around, and ask, “Why aren’t you…?”

by Dwayne Phillips

The frustrating circle of measuring measuring.

The most difficult photograph to take is of my camera. I need another camera to take a photograph of my camera. I don’t want a poor-quality photograph of my camera, so I need a second camera that is just as good or better than my camera.

If I had a second camera that is just as good or better than my camera, it would be my camera and I would discard my first camera. Now I need another camera that is…I think we are going around in circles.

Hence, we are stuck with the problem of measuring a measuring device. How do we know the second measuring device is as good or better than the first measuring device. If it is better, we must replace the first measuring device and…wait, we are back in that circle, again.

There must be a better answer. I don’t know what it is, but I would like to believe there is a better answer. All I wanted was a picture of my camera.

This science stuff isn’t so easy.

by Dwayne Phillips

Are you a data scientist? Have you ever read that old green textbook? Probably not, and that is unfortunate.

I’ve been delving into data science lately. I’ve applied for a few jobs with the title “Data Scientist.” I’m not sure what a data scientist is, but much of what I have learned about the profession I learned a few decades ago in a green textbook.

The book, of course as old folks know, is “Pattern Classification and Scene Analysis” by Richard Duda and Peter Hart. I actually found a place where you can buy the first edition.

Old? Yes. Nevertheless, P(H|E) still equals (P(E|H)*P(H))/P(E). That is somewhat of an inconvenient truth, but that is the way it is.

We move onward as Intel creates yet another generation of processors with more multiplications per second at a lower $ per everything. With this, n becomes larger, hardware budgets fall, and some of those equations become possible. That is, of course, if we understand those equations in the back of that green hardback book.

In my opinion—this is, after all, my blog—you cannot be a data scientist without knowing who Duda and Hart are and what they did. Then again, I am old—apparently too old to be hired in such a new profession as Data Scientist.

by Dwayne Phillips

One basic progression of understanding.

I understand this well enough to:

• understand it in my mind
• use it
• explain it clearly to others
• invent it
• invent what comes next
• ignore it