Evidence

Format Web Page

"Croatia’s futuristic open-road tolling system will use Automatic License Plate Recognition alongside ENC technology."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE Can Our Future Look Like the Past?

Source title O sustavu CROLIBERTAS

Source authors Hrvatske autoceste

Source year 2026

Confidence high

Importance mid

Format Web Page

"iRobot hand-sculpted hundreds of fake dog turds out of Play-Doh and paint to provide training examples for its Roomba object-recognition system."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE Can Our Future Look Like the Past?

Source title With New Roomba j7, iRobot Wants to Understand Our Homes

Source authors Evan Ackerman

Source year 2021

Confidence high

Importance mid

Format Web Page

"Roomba uses onboard computer vision to recognise floor-level obstacles such as socks, cables and dog poop."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE Can Our Future Look Like the Past?

Source title With New Roomba j7, iRobot Wants to Understand Our Homes

Source authors Evan Ackerman

Source year 2021

Confidence high

Importance mid

Format Online paper / PDF

"When image classifiers encounter objects in unfamiliar contexts, their learned shortcuts can cause unexpected recognition failures."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Fighting Fire with Fire: Avoiding DNN Shortcuts through Priming

Source authors Chuan Wen, Jianing Qian, Jierui Lin, Jiaye Teng, Dinesh Jayaraman, Yang Gao

Source year 2022

Confidence high

Importance high

Format Online paper / PDF

"Models can learn patterns that correlate with training labels rather than the intended underlying concepts, producing shortcut solutions that fail under unfamiliar conditions."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Shortcut Learning in Deep Neural Networks

Source authors Robert Geirhos, Jörn-Henrik Jacobsen, Claudio Michaelis, Richard Zemel, Wieland Brendel, Matthias Bethge, Felix A. Wichmann

Source year 2020

Confidence high

Importance high

Format Web Page

"If most training images of cows show green grass, a model may learn to associate the label “cow” with the background instead of relying primarily on the animal itself."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Avoiding shortcut solutions in artificial intelligence

Source authors Adam Zewe

Source year 2021

Confidence high

Importance high

Format Online paper / PDF

"Deep learning constructs internal representations of patterns that predict labels such as “cow” from large collections of training images."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Deep learning

Source authors Yann LeCun, Yoshua Bengio, Geoffrey Hinton

Source year 2015

Confidence high

Importance high

Format Online paper / PDF

"In traditional machine learning, engineers often had to define or extract the features that might matter, while deep learning largely automates the construction of internal representations."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Deep learning

Source authors Yann LeCun, Yoshua Bengio, Geoffrey Hinton

Source year 2015

Confidence high

Importance high

Format Online paper / PDF

"ANNs are, at base, “just” stacked mathematical functions, capable of approximating complex relationships."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Multilayer feedforward networks are universal approximators

Source authors Kurt Hornik, Maxwell Stinchcombe, Halbert White

Source year 1989

Confidence high

Importance high

Format Online paper / PDF

"The early 2000s hosted a fiery argument over the role of statistics and algorithmic modelling, another Taste War that would follow machine learning for years."

View Source

Location PART 3 - THE THIRD AXIS CHAPTER 13: SPECKLED INTELLIGENCE “Just” Statistics

Source title Statistical Modeling: The Two Cultures

Source authors Leo Breiman

Source year 2001

Confidence high

Importance high