v3.9.0

RxInfer v3.9.0

Diff since v3.8.4

Merged pull requests:

• Add Binomial Regression Tests and Example Notebook (#403) (@ismailsenoz)
• Move examples out of the repository (#404) (@bvdmitri)

Closed issues:

• Broken example on rxinfer website (#401)

v3.8.4

RxInfer v3.8.4

Diff since v3.8.3

v3.8.3

RxInfer v3.8.3

Diff since v3.8.2

Merged pull requests:

• Create a separate CI action for formatting (#307) (@bvdmitri)
• add structural dynamics example (#392) (@vflores-io)
• Update README.md (#398) (@albertpod)

v3.8.2

RxInfer v3.8.2

Diff since v3.8.1

Merged pull requests:

• Add sharp bits section and invite users to the discussions upon error (#391) (@bvdmitri)

Closed issues:

• An overview of "sharp bits" of RxInfer (#60)
• Misleading mean-field constraint comment in HGF demo (#383)
• Intercept errors and point users to a discussion platform (#385)
• Enhance error handling: implement user-friendly error messages for common issues (#386)

v3.8.1

RxInfer v3.8.1

Diff since v3.8.0

Merged pull requests:

• Fix HGF notebook (#387) (@HoangMHNguyen)
• Dev node in a rule (#394) (@bvdmitri)
• Use GraphPPL 4.5.0 (#396) (@bvdmitri)

Closed issues:

• Issue in the dot product within a NormalMeanVariance distribution (#390)

v3.8.0

RxInfer v3.8.0

Diff since v3.7.2

Merged pull requests:

• Added possible solution to constrain form error print (#370) (@wmkouw)
• Update deps to GraphPPL 4.4.0 and update docs on the visualization (#384) (@bvdmitri)

Closed issues:

• RuleMethodError: improve debugging information (#366)

v3.7.2

RxInfer v3.7.2

Diff since v3.7.1

Merged pull requests:

• Fix tests for new version of Julia (#369) (@bvdmitri)
• Bump RMP version (#375) (@bvdmitri)
• Change readme to address numfocus comments (#377) (@ismailsenoz)
• add example: LitterModel (#378) (@LearnableLoopAI)
• fix(cviprojection-tests): prjparams is removed from CVIProjection (#381) (@Nimrais)

Closed issues:

• Add tags to the description of the repository (#374)
• CVI projection broken tests (#380)

v3.7.1

RxInfer v3.7.1

Diff since v3.7.0

Merged pull requests:

• CompatHelper: bump compat for Static to 1, (keep existing compat) (#361) (@github-actions[bot])
• Simplify model definition in kalman filter example (#362) (@wouterwln)
• Restore error message for unparametrized Normal/MvNormal/Gamma (#364) (@bvdmitri)
• Document := operator better (#365) (@bvdmitri)

Closed issues:

• := operator is not mentioned in RxInfer docs, leading to confusion (#359)
• x ~ Normal(0, 1) should error and hinting users to use Normal(mean = ..., var/prec = ...) (#360)

v3.7.0

RxInfer v3.7.0

Diff since v3.6.1

Merged pull requests:

• Add custom node contraction (#342) (@blolt)
• Add UnfactorizedData (#356) (@wouterwln)

Closed issues:

• Allow contraction of composite nodes (#287)
• Predictions in the Bike Rental Demand demo are worse than in RxInfer v2 due to factorization constraints (#291)
• Initialization sometimes causes exponential scaling in runtime (#353)
• Prediction of data variables when some are missing (#355)
• Linear model with log-link on sigma (exp(sigma)) errors: are there plans to support it in the future? (#357)

v3.6.1

RxInfer v3.6.1

Diff since v3.6.0

Merged pull requests:

• Update README.md to include external contributors (#348) (@albertpod)
• Update GitHub Actions to use upload-artifact@v4 (#352) (@bvdmitri)
• Add MultiAgentTrajectoryPlanning to tests (#354) (@bvdmitri)