Skip to content

Tags: Julia-XAI/ExplainableAI.jl

Tags

v0.10.3

[Diff since v0.10.2](v0.10.2...v0.10.3)

v0.10.2

[Diff since v0.10.1](v0.10.1...v0.10.2)

**Merged pull requests:**
- Tested GPU support for `Gradient`, `InputTimesGradient`, `SmoothGrad`, `IntegratedGradients` (#184) (@adrhill)

v0.10.1

[Diff since v0.10.0](v0.10.0...v0.10.1)

**Merged pull requests:**
- Fix bug in `NoiseAugmentation` constructor, improve docs (#183) (@adrhill)

v0.10.0

[Diff since v0.9.0](v0.9.0...v0.10.0)

- ![BREAKING][badge-breaking] Fix default noise level for `NoiseAugmentation` ([#179])
- ![Enhancement][badge-enhancement] Constant memory input augmentations ([#180])

[#180]: #180
[#179]: #179
[badge-breaking]: https://img.shields.io/badge/BREAKING-red.svg
[badge-enhancement]: https://img.shields.io/badge/enhancement-blue.svg

**Merged pull requests:**
- Bump codecov/codecov-action from 4 to 5 (#178) (@dependabot[bot])
- Fix default noise level for `NoiseAugmentation` (#179) (@adrhill)
- Constant memory input augmentations (#180) (@adrhill)
- Update DifferentiationInterface dependency (#181) (@adrhill)
- Remove heatmapping docs (#182) (@adrhill)

v0.9.0

[Diff since v0.8.1](v0.8.1...v0.9.0)

- ![Feature][badge-feature] Support selection of AD backends via DifferentiationInterface.jl. ([#167])
- ![BREAKING][badge-breaking] For gradient-based XAI methods, an AD backend must now be manually loaded. To keep using the default Zygote backend, this simply requires adding `using Zygote` to your code. ([#177])
- ![BREAKING][badge-breaking] `Gradient`, `InputTimesGradient` and `GradCAM` analyzers now have an additional `backend` field and type parameter. ([#167])
- ![BREAKING][badge-breaking] Update XAIBase interface to `v4`. This adds a field to the `Explanation` return type and removes the `add_batch_dim` keyword argument. Refer to the [XAIBase.jl changelog](https://github.com/Julia-XAI/XAIBase.jl/blob/main/CHANGELOG.md#version-v400) for more information. ([#174])

[#177]: #177
[#174]: #174
[#167]: #167
[badge-breaking]: https://img.shields.io/badge/BREAKING-red.svg
[badge-deprecation]: https://img.shields.io/badge/deprecation-orange.svg
[badge-feature]: https://img.shields.io/badge/feature-green.svg

**Merged pull requests:**
- Update XAIBase dependency to v4.0.0 (#166) (@adrhill)
- Use DifferentiationInterface for gradient-based analyzers (#167) (@adrhill)
- Update CI workflows (#170) (@adrhill)
- Update CI: Give CompatHelper permission to write PRs (#172) (@adrhill)
- Update CompatHelper v2 (#173) (@adrhill)
- Drop compatibility with pre-LTS versions (#174) (@adrhill)
- Update benchmarks and test them with PkgJogger (#175) (@adrhill)
- Add full example to README (#176) (@adrhill)
- Remove Zygote from dependencies (#177) (@adrhill)

**Closed issues:**
- Use AbstractDifferentiation.jl (#7)
- Allow selection of AD backends via DifferentiationInterface (#165)
- Update benchmarks (#168)
- Incomplete starter code (#169)
- Update benchmarks and test them with PkgJogger (#171)

v0.8.1

[Diff since v0.8.0](v0.8.0...v0.8.1)

**Merged pull requests:**
- Test type stability with JET.jl (#163) (@adrhill)

v0.8.0

[Diff since v0.7.0](v0.7.0...v0.8.0)

This release removes the automatic reexport of heatmapping functionality.
Users are now required to manually load [VisionHeatmaps.jl][VisionHeatmaps] and/or [TextHeatmaps.jl][TextHeatmaps].

This reduces the maintenance burden for new heatmapping features and the amount of dependencies for users who don't require heatmapping functionality.

* ![BREAKING][badge-breaking] Removed reexport of heatmapping functionality by updating XAIBase dependency ([#162][pr-162]).
* ![Feature][badge-feature] Added `GradCAM` analyzer ([#155][pr-155]). Try it with [VisionHeatmaps.jl][VisionHeatmaps]'s new `heatmap_overlay` feature.

[VisionHeatmaps]: https://julia-xai.github.io/XAIDocs/VisionHeatmaps/stable/
[TextHeatmaps]: https://julia-xai.github.io/XAIDocs/TextHeatmaps/stable/

[pr-162]: #162
[pr-155]: #155
[badge-breaking]: https://img.shields.io/badge/BREAKING-red.svg
[badge-feature]: https://img.shields.io/badge/feature-green.svg

**Merged pull requests:**
- Add GradCAM analyzer (#155) (@JeanAnNess)
- Bump codecov/codecov-action from 3 to 4 (#161) (@dependabot[bot])
- Update XAIBase to v3.0.0 and document `heatmap_overlay` (#162) (@adrhill)

v0.7.0

[Diff since v0.6.3](v0.6.3...v0.7.0)

This release moves the core interface (`Explanation`, `heatmap`, `analyze`) into a separate package called [XAIBase.jl](https://github.com/Julia-XAI/XAIBase.jl). Developers can make use of the [XAIBase.jl interface](https://julia-xai.github.io/XAIDocs/XAIBase/) to quickly implement or prototype new methods without having to write boilerplate code.

As announced with version `v0.6.2`, this is first release without LRP, which has been moved to a separate package called [RelevancePropagation.jl](https://github.com/Julia-XAI/RelevancePropagation.jl). This separation is enabled by the new common XAIBase.jl interface.

- ![BREAKING][badge-breaking] Move core interface into XAIBase.jl package ([#154][pr-154]).
  - Renamed `Explanation` field `neuron_selection` to `output_selection`
  - Added `Explanation` field `heatmap` for heatmapping presets
- ![BREAKING][badge-breaking] Move LRP into RelevancePropagation.jl ([#157][pr-157])
- ![BREAKING][badge-breaking] Remove ImageNet preprocessing utilities ([#159][pr-159])
- ![Documentation][badge-docs] Partially move documentation into the [Julia-XAI ecosystem documentation](https://github.com/Julia-XAI/XAIDocs)

[badge-breaking]: https://img.shields.io/badge/BREAKING-red.svg
[badge-deprecation]: https://img.shields.io/badge/deprecation-orange.svg
[badge-feature]: https://img.shields.io/badge/feature-green.svg
[badge-enhancement]: https://img.shields.io/badge/enhancement-blue.svg
[badge-bugfix]: https://img.shields.io/badge/bugfix-purple.svg
[badge-security]: https://img.shields.io/badge/security-black.svg
[badge-experimental]: https://img.shields.io/badge/experimental-lightgrey.svg
[badge-maintenance]: https://img.shields.io/badge/maintenance-gray.svg
[badge-docs]: https://img.shields.io/badge/docs-orange.svg

[pr-159]: #159
[pr-157]: #157
[pr-154]: #154

**Merged pull requests:**
- Move core interface into XAIBase.jl package (#154) (@adrhill)
- Bump actions/cache from 3 to 4 (#156) (@dependabot[bot])
- Move LRP into RelevancePropagation.jl (#157) (@adrhill)
- Remove ImageNet preprocessing utilities (#159) (@adrhill)
- Remove unused dependencies and utils (#160) (@adrhill)

**Closed issues:**
- LRP rule coverage for Flux layers (#9)
- Add visualization method for language models (#79)
- Add XAIBase dependency (#151)
- Move LRP into separate package (#152)
- Remove ImageNet preprocessing code (#153)
- Update README before 1.0 release (#158)

v0.6.3

[Diff since v0.6.2](v0.6.2...v0.6.3)

**Merged pull requests:**
- Allow Gradient analyzers on non-Flux models (#150) (@adrhill)

v0.6.2

[Diff since v0.6.1](v0.6.1...v0.6.2)

**Merged pull requests:**
- Add Concept Relevance Propagation (#146) (@adrhill)
- Fix `EpsilonPlusFlat` and `EpsilonAlpha2Beta1Flat` composites (#147) (@adrhill)
- Document CRP (#148) (@adrhill)
- Transfer repository to Julia-XAI organization (#149) (@adrhill)