← All workshopsRead Learning the Metric Itself

Computer Vision with Neural Networks

Make a useful difference count

BOTIE

Return anchor-to-match squared distance minus anchor-to-nonmatch squared distance.
Ready to test
0 passed · 0 failed · 1 not run

An empty floor. Start with the information your machine needs.

Received result

Your answer will arrive here.

Examples, experiments and workspace controls
Live values

Connect a tool's inputs to see what it produces. You can try a branch before connecting it to Result.

Try your own inputs

Use a copy of the example to try your own data. Keep the problem's assumptions in mind, and predict the effect of a change before checking the output. You can also add another branch to compare two ways of doing the calculation.

Start at an output arrow, then click or drag to an input. Drag empty space to pan. Delete removes a selection. Ctrl/Cmd+Z undoes a change.
Context for this challenge

Return anchor-to-match squared distance minus anchor-to-nonmatch squared distance.

Pictures that should match do not necessarily start close together in a representation. We have one picture, a matching example, and a nonmatching example. Let's measure whether the matching one is closer by the required margin. Training would use that error to change the representation.

Inspect the component calculation
Subtract(
  Starting value: Add a list(
    Numbers: Square(
      Value or list: Subtract(
        Starting value: Anchor vector
        Value to remove: Matching vector
      )
    )
  )
  Value to remove: Add a list(
    Numbers: Square(
      Value or list: Subtract(
        Starting value: Anchor vector
        Value to remove: Nonmatching vector
      )
    )
  )
)
Examples and results
Anchor vector
  1. 00
  2. 10
Matching vector
  1. 01
  2. 10
Nonmatching vector
  1. 00.5
  2. 10
Required result
0.75
Try a fresh construction

This clears this draft, including its undo history. Earlier successful attempts stay recorded. Build again without opening hints to record an independent attempt.

Stuck? Inspect a worked construction

The tools below form one possible solution. You can assemble it, follow its signals, and undo that change to return to your own construction.

Assumptions and test precision

This is a small construction problem. The lesson explains the assumptions and the parts of the full method that this example leaves out. Numerical tests allow a difference of 0.001. Lists and structures must also have the right shape.

Your last tested pieces can start the full construction. An existing draft there keeps its own pieces.

The nonmatching picture shares the anchor's vector

This check uses the pieces from “Measure a violated triplet margin”.

BOTIE

A nonmatching picture currently has exactly the anchor's representation. Compute the triplet loss using squared Euclidean distances and a margin of one. Return the loss after the nonnegative cutoff.
Anchor vector
  1. 01
  2. 11
Matching vector
  1. 02
  2. 11
Nonmatching vector
  1. 01
  2. 11

Use a number, text, or an array matching the requested result. For a matrix, put each row in brackets. Saving your prediction runs your current construction on these inputs and checks the earlier examples too.

Build a construction in the workspace above first. This case will test those pieces.

Does a zero loss on one triplet establish that all matching pictures are close?

A metric-learning objective must be evaluated across suitable examples; computing it does not itself train the representation.

This is a small construction problem. The lesson explains the assumptions and the parts of the full method that this example leaves out.

Work through the Python challenges →
Continue the courseSearching a Collection of Pictures →
← A Vector for a Picture