Robot Learning with a Biologically-Inspired Brain (BECCA)

Brandon Rohrer · Beginner ·📰 AI News & Updates ·17y ago

Key Takeaways

Demonstrates robot learning using a biologically-inspired brain architecture called BECCA

Full Transcript

[Music] directed robots run by themselves but can be trained like dogs they learn the meaning of commands by getting rewards in the directed robots Lab at Sandy National laborator stories a surveyor srv1 robots named Ral is learning to hide and seek hide means to find a picture that is either all light or all dark for him seek means to find a picture with lots of light and dark when he first starts up Ral doesn't know how moving will change what he sees [Music] he just tries different movements and remembers what happens each time eventually he learns that moving forward toward the wall is a good way to hide and that backing up is usually a good way to seek sometimes roal gets bored or curious and tries something different this helps him figure out the best way to do what he is trying to do most robots know a lot about themselves like the size of the room they're in and whether they have wheels or arms Ralph doesn't know any of that he doesn't even know that his eye is an eye and not a nose he doesn't know how far he'll turn when he spins he only remembers what his sensors tell him and what rewards he gets for hiding and seeking ral's brain is a computer program called s learning s learning doesn't care what kind of robot it is connected to it can be a brain for lots of different kinds of robots s learning learns whatever it gets rewarded for like a baby s learning uses all its past experiences whenever it has to learn to do something new it might even be able to help a robot do hard things like run on two legs or learn to talk but for now Ralph is still learning kind of fetch [Music]

Original Description

Code and documentation: https://github.com/brohrer/ BECCA users group: https://groups.google.com/forum/?fromgroups#!forum/becca_users BECCA community: http://www.openbecca.org A tracked robot learns some of the basics of its environment by exploring. It software "brain" is BECCA, a brain-emulating cognition and control architecture. BECCA gives the robot the ability to learn from its experience and to develop very simple problem solving strategies. Video released as SAND report # 2009-2734 P.
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Playlist

Uploads from Brandon Rohrer · Brandon Rohrer · 1 of 60

← Previous Next →
Robot Learning with a Biologically-Inspired Brain (BECCA)
Robot Learning with a Biologically-Inspired Brain (BECCA)
Brandon Rohrer
2 BECCA talk at AGI 2011
BECCA talk at AGI 2011
Brandon Rohrer
3 Robot Learning with a Biologically-Inspired Brain (BECCA), The Sequel
Robot Learning with a Biologically-Inspired Brain (BECCA), The Sequel
Brandon Rohrer
4 BECCA listens to The Hobbit
BECCA listens to The Hobbit
Brandon Rohrer
5 Learning the building blocks of speech: BECCA extracts a hierarchy of audio features
Learning the building blocks of speech: BECCA extracts a hierarchy of audio features
Brandon Rohrer
6 BECCA listens for sound effects in The Hobbit
BECCA listens for sound effects in The Hobbit
Brandon Rohrer
7 BECCA finds movie trailers while watching the Big Bang Theory
BECCA finds movie trailers while watching the Big Bang Theory
Brandon Rohrer
8 Listening for unexpected sounds: BECCA detects anomalies in audio data
Listening for unexpected sounds: BECCA detects anomalies in audio data
Brandon Rohrer
9 Learning the building blocks of vision: BECCA extracts a spatio-temporal hierarchy of features
Learning the building blocks of vision: BECCA extracts a spatio-temporal hierarchy of features
Brandon Rohrer
10 Watching for the unexpected: BECCA detects anomalies in video data
Watching for the unexpected: BECCA detects anomalies in video data
Brandon Rohrer
11 BECCA finds a stationary target
BECCA finds a stationary target
Brandon Rohrer
12 BECCA finds a stationary target at 3X speed
BECCA finds a stationary target at 3X speed
Brandon Rohrer
13 BECCA watches the X-men and Bruce Lee
BECCA watches the X-men and Bruce Lee
Brandon Rohrer
14 BECCA plays Quidditch
BECCA plays Quidditch
Brandon Rohrer
15 BECCA chases a ball
BECCA chases a ball
Brandon Rohrer
16 BECCA chases a ball, part 2
BECCA chases a ball, part 2
Brandon Rohrer
17 Becca chases a ball, part 3
Becca chases a ball, part 3
Brandon Rohrer
18 BECCA creates features from MNIST
BECCA creates features from MNIST
Brandon Rohrer
19 How reinforcement learning works in Becca 7
How reinforcement learning works in Becca 7
Brandon Rohrer
20 Deep Learning Demystified
Deep Learning Demystified
Brandon Rohrer
21 How Data Science Works
How Data Science Works
Brandon Rohrer
22 How Convolutional Neural Networks work
How Convolutional Neural Networks work
Brandon Rohrer
23 How Bayes Theorem works
How Bayes Theorem works
Brandon Rohrer
24 How Deep Neural Networks Work
How Deep Neural Networks Work
Brandon Rohrer
25 Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM)
Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM)
Brandon Rohrer
26 How Support Vector Machines work / How to open a black box
How Support Vector Machines work / How to open a black box
Brandon Rohrer
27 How autocorrelation works
How autocorrelation works
Brandon Rohrer
28 Getting closer to human intelligence through robotics
Getting closer to human intelligence through robotics
Brandon Rohrer
29 A minimalist's guide to slicing and indexing pandas DataFrames
A minimalist's guide to slicing and indexing pandas DataFrames
Brandon Rohrer
30 How decision trees work
How decision trees work
Brandon Rohrer
31 Data scientist archetypes
Data scientist archetypes
Brandon Rohrer
32 How to use python's datetime package
How to use python's datetime package
Brandon Rohrer
33 How optimization for machine learning works, part 1
How optimization for machine learning works, part 1
Brandon Rohrer
34 How optimization for machine learning works, part 2
How optimization for machine learning works, part 2
Brandon Rohrer
35 How optimization for machine learning works, part 3
How optimization for machine learning works, part 3
Brandon Rohrer
36 How optimization for machine learning works, part 4
How optimization for machine learning works, part 4
Brandon Rohrer
37 How convolutional neural networks work, in depth
How convolutional neural networks work, in depth
Brandon Rohrer
38 How to pick a machine learning model 4: Splitting the data
How to pick a machine learning model 4: Splitting the data
Brandon Rohrer
39 How to pick a machine learning model 3: Choosing a loss function
How to pick a machine learning model 3: Choosing a loss function
Brandon Rohrer
40 How to pick a machine learning model 2: Separating signal from noise
How to pick a machine learning model 2: Separating signal from noise
Brandon Rohrer
41 How to pick a machine learning model 1: Choosing between models
How to pick a machine learning model 1: Choosing between models
Brandon Rohrer
42 How to pick a machine learning model 5: Navigating assumptions
How to pick a machine learning model 5: Navigating assumptions
Brandon Rohrer
43 What do neural networks learn?
What do neural networks learn?
Brandon Rohrer
44 Interview with iRobot's Director of Data Science Angela Bassa
Interview with iRobot's Director of Data Science Angela Bassa
Brandon Rohrer
45 How Backpropagation Works
How Backpropagation Works
Brandon Rohrer
46 Evolutionary Powell's method: A discrete optimizer for hyperparameter optimization
Evolutionary Powell's method: A discrete optimizer for hyperparameter optimization
Brandon Rohrer
47 1D convolution for neural networks, part 1: Sliding dot product
1D convolution for neural networks, part 1: Sliding dot product
Brandon Rohrer
48 1D convolution for neural networks, part 2: Convolution copies the kernel
1D convolution for neural networks, part 2: Convolution copies the kernel
Brandon Rohrer
49 1D convolution for neural networks, part 3: Sliding dot product equations longhand
1D convolution for neural networks, part 3: Sliding dot product equations longhand
Brandon Rohrer
50 1D convolution for neural networks, part 4: Convolution equation
1D convolution for neural networks, part 4: Convolution equation
Brandon Rohrer
51 1D convolution for neural networks, part 5: Backpropagation
1D convolution for neural networks, part 5: Backpropagation
Brandon Rohrer
52 1D convolution for neural networks, part 6: Input gradient
1D convolution for neural networks, part 6: Input gradient
Brandon Rohrer
53 1D convolution for neural networks, part 7: Weight gradient
1D convolution for neural networks, part 7: Weight gradient
Brandon Rohrer
54 1D convolution for neural networks, part 8: Padding
1D convolution for neural networks, part 8: Padding
Brandon Rohrer
55 1D convolution for neural networks, part 9: Stride
1D convolution for neural networks, part 9: Stride
Brandon Rohrer
56 The Four Grand Challenges of Robots in the Home
The Four Grand Challenges of Robots in the Home
Brandon Rohrer
57 How Convolution Works
How Convolution Works
Brandon Rohrer
58 The Softmax neural network layer
The Softmax neural network layer
Brandon Rohrer
59 Batch normalization
Batch normalization
Brandon Rohrer
60 Getting ready to learn Python, Mac edition #1: Files and directories
Getting ready to learn Python, Mac edition #1: Files and directories
Brandon Rohrer

Related Reads

📰
Claude Sonnet 5 Didn’t Just Get Smarter. It Changed the Economics of AI.
Learn how Claude Sonnet 5's advancements changed the economics of AI, making 'good enough AI' viable for production, and understand the implications for AI development and deployment
Medium · AI
📰
Claude Sonnet 5 Didn’t Just Get Smarter. It Changed the Economics of AI.
Claude Sonnet 5's improved AI capabilities have transformed the economics of AI, making it more viable for production
Medium · Machine Learning
📰
The AI Career Toolkit That Replaced My Job Hunt in 2026
Learn how to leverage AI tools to enhance your job search and career development in 2026
Dev.to · freelancewith_ai
📰
The AI Problem Nobody Saw Coming: The Decline Of Curiosity And Meaning
The rise of AI may lead to a decline in human curiosity, affecting innovation and our sense of meaning, and it's crucial to understand this potential consequence
Forbes Innovation
Up next
Man dies after horror Gold Coast house fire; high-speed Sydney motorway pursuit | 9 News Australia
9 News Australia
Watch →