[MINI] Feed Forward Neural Networks
Data Skeptic24 Mar 2017

[MINI] Feed Forward Neural Networks

Feed Forward Neural Networks

In a feed forward neural network, neurons cannot form a cycle. In this episode, we explore how such a network would be able to represent three common logical operators: OR, AND, and XOR. The XOR operation is the interesting case.

Below are the truth tables that describe each of these functions.

AND Truth Table Input 1 Input 2 Output 0 0 0 0 1 0 1 0 0 1 1 1 OR Truth Table Input 1 Input 2 Output 0 0 0 0 1 1 1 0 1 1 1 1 XOR Truth Table Input 1 Input 2 Output 0 0 0 0 1 1 1 0 1 1 1 0

The AND and OR functions should seem very intuitive. Exclusive or (XOR) if true if and only if exactly single input is 1. Could a neural network learn these mathematical functions?

Let's consider the perceptron described below. First we see the visual representation, then the Activation function , followed by the formula for calculating the output.

Can this perceptron learn the AND function?

Sure. Let and

What about OR?

Yup. Let and

An infinite number of possible solutions exist, I just picked values that hopefully seem intuitive. This is also a good example of why the bias term is important. Without it, the AND function could not be represented.

How about XOR?

No. It is not possible to represent XOR with a single layer. It requires two layers. The image below shows how it could be done with two laters.

In the above example, the weights computed for the middle hidden node capture the essence of why this works. This node activates when recieving two positive inputs, thus contributing a heavy penalty to be summed by the output node. If a single input is 1, this node will not activate.

Universal approximation theorem tells us that any continuous function can be tightly approximated using a neural network with only a single hidden layer and a finite number of neurons. With this in mind, a feed forward neural network should be adaquet for any applications. However, in practice, other network architectures and the allowance of more hidden layers are empirically motivated.

Other types neural networks have less strict structal definitions. The various ways one might relax this constraint generate other classes of neural networks that often have interesting properties. We'll get into some of these in future mini-episodes.

Check out our recent blog post on how we're using Periscope Data cohort charts.

Thanks to Periscope Data for sponsoring this episode. More about them at periscopedata.com/skeptics

Episoder(588)

What the Antlion Knows

What the Antlion Knows

Our guest is Becky Hansis-O'Neil, a Ph.D. student at the University of Missouri, St Louis, and our co-host for the new "Animal Intelligence" season. Becky shares her background on how she got into the field of behavioral intelligence and biology.

30 Jan 202441min

AI Roundtable

AI Roundtable

Kyle is joined by friends and former guests Pramit Choudhary and Frank Bell to have an open discussion of the impacts LLMs and machine learning have had in the past year on industry, and where things may go in the current year.

17 Jan 202450min

Uncontrollable AI Risks

Uncontrollable AI Risks

We are joined by Darren McKee, a Policy Advisor and the host of Reality Check — a critical thinking podcast. Darren gave a background about himself and how he got into the AI space. Darren shared his thoughts on AGI's achievements in the coming years. He defined AGI and discussed how to differentiate an AGI system. He also shared whether AI needs consciousness to be AGI. Darren discussed his worry about AI surpassing human understanding of the universe and potentially causing harm to humanity. He also shared examples of how AI is already used for nefarious purposes. He explored whether AI possesses inherently evil intentions and gave his thoughts on regulating AI.

27 Des 202338min

I LLM and You Can Too

I LLM and You Can Too

It took a massive financial investment for the first large language models (LLMs) to be created. Did their corporate backers lock these tools away for all but the richest? No. They provided comodity priced API options for using them. Anyone can talk to Chat GPT or Bing. What if you want to go a step beyond that and do something programatic? Kyle explores your options in this episode.

23 Des 202323min

Q&A with Kyle

Q&A with Kyle

We celebrate episode 1000000000 with some Q&A from host Kyle Polich. We boil this episode down to four key questions: 1) How do you find guests 2) What is Data Skeptic all about? 3) What is Kyle all about? 4) What are Kyle's thoughts on AGI? Thanks to our sponsorsdataannotation.tech/programmers https://www.webai.com/dataskeptic

19 Des 202340min

LLMs for Data Analysis

LLMs for Data Analysis

In this episode, we are joined by Amir Netz, a Technical Fellow at Microsoft and the CTO of Microsoft Fabric. He discusses how companies can use Microsoft's latest tools for business intelligence. Amir started by discussing how business intelligence has progressed in relevance over the years. Amir gave a brief introduction into what Power BI and Fabric are. He also discussed how Fabric distinguishes from other BI tools by building an end-to-end tool for the data journey. Amir spoke about the process of building and deploying machine learning models with Microsoft Fabric. He shared the difference between Software as a Service (SaaS) and Platform as a Service (PaaS). Amir discussed the benefits of Fabric's auto-integration and auto-optimization abilities. He also discussed the capabilities of Copilot in Fabric. He also discussed exciting future developments planned for Fabric. Amir shared techniques for limiting Copilot hallucination.

12 Des 202329min

AI Platforms

AI Platforms

Our guest today is Eric Boyd, the Corporate Vice President of AI at Microsoft. Eric joins us to share how organizations can leverage AI for faster development. Eric shared the benefits of using natural language to build products. He discussed the future of version control and the level of AI background required to get started with Azure AI. He mentioned some foundational models in Azure AI and their capabilities. Follow Eric on LinkedIn to learn more about his work. Visit today's sponsor at https://webai.com/dataskeptic

4 Des 202333min

Deploying LLMs

Deploying LLMs

We are excited to be joined by Aaron Reich and Priyanka Shah. Aaron is the CTO at Avanade, while Priyanka leads their AI/IoT offering for the SEA Region. Priyanka is also the MVP for Microsoft AI. They join us to discuss how LLMs are deployed in organizations.

27 Nov 202335min

Populært innen Vitenskap

fastlegen
rekommandert
jss
rss-rekommandert
tingenes-tilstand
sinnsyn
rss-nysgjerrige-norge
villmarksliv
dekodet-2
forskningno
doktor-fives-podcast
rss-paradigmepodden
vett-og-vitenskap-med-gaute-einevoll
pod-britannia
psykopoden
tidlose-historier
diagnose
tomprat-med-gunnar-tjomlid
nevropodden
abid-nadia-skyld-og-skam