[Hamidreza Keshavarz (hamidkm9)]

Why are these changes needed?

FLAML's cheap linear slot for classification is lrl1, and a single
LogisticRegression(saga) fit is not interruptible, so a small time_budget
can be overrun by orders of magnitude (see #1612). This PR adds SEFR, a linear-time, closed-form classifier, as another
low-cost option.

SEFR computes one weight per feature plus a bias from class-conditional feature
means. Fitting is a closed form with no iterative optimization, and it takes three
passes over the data whatever the number of classes: the feature range, the
per-class feature sums (one matrix product), and the spread of the training
scores. Beyond the scaled input, fitting needs O(n_samples + n_classes * n_features)
auxiliary memory plus bounded score chunks.
Each binary SEFR head is n_features + 1 floats; the fitted classifier also
stores per-feature scaling parameters, one head per class for multiclass targets,
and the calibration (one scale, plus one bias per class).

Keshavarz, Saniee Abadeh, Rawassizadeh (2020),
"SEFR: A Fast Linear-Time Classifier for Ultra-Low Power Devices", arXiv:2006.04620

This PR registers two estimators. Both are opt-in and left out of
default_estimator_list, the same way histgb, kneighbor and svc are:

• "sefr" is SEFR itself. Feature scaling is part of the estimator, since
  SEFR's weight formula needs non-negative features and FLAML does not scale
  anywhere in its pipeline. Dense input is min-max scaled and clipped to [0, 1]
  at predict time; sparse input is scaled by its column max and rejected if it
  has negative entries. Margins are calibrated because FLAML's default
  multiclass metric is log_loss and raw SEFR emits margins, not probabilities.
  Margins map to calibrated logits a * margin + b (one shared scale, one bias
  per head) and to their sigmoid or softmax, so predict, decision_function
  and predict_proba always agree. The default "sigmoid" is closed form;
  "platt" fits a and b by maximum likelihood and is in the search space.
  class_weight acts as misclassification costs on this calibrated decision.
  On iris (5-fold) log loss is 0.908 uncalibrated, 0.555 with sigmoid and
  0.209 with platt.
• "sefr_boost" is AdaBoost over SEFR base learners, which gives the search a
  real space that is still cheap to explore. It also tunes the base learner's
  calibration, starting at "platt".

This adds no dependencies. The algorithm is a handful of NumPy operations, and
FLAML already requires NumPy.

Budget adherence

Max wall-clock / requested time_budget over budgets of 1, 2, 5 and 10 seconds,
on the classification datasets in test/default/all/metafeatures.csv
(pandas 2.3.3, scikit-learn 1.5.2):

dataset	n_train	n_feat	n_cls	lrl1	lgbm	sefr	sefr_boost
car	1,296	6	2	1.00x	1.02x	1.00x	1.03x
dilbert	7,500	2,000	5	24.8x	1.74x	1.54x	1.56x
Amazon_employee_access	24,576	9	2	1.14x	1.67x	1.01x	1.41x
adult	36,631	14	2	1.25x	1.71x	1.01x	1.13x
connect-4	50,667	126	3	8.47x	1.54x	1.05x	1.15x
Dionis	312,141	60	355	770x	3.33x	1.68x	18.1x
Albert	318,930	78	2	10.3x	1.21x	1.12x	2.70x
Airlines	404,537	7	2	5.16x	1.10x	1.04x	1.43x
poker	768,757	10	2	5.51x	1.24x	1.04x	1.45x

sefr stays within 1.12x on 7 of 9 datasets and within 1.7x on all 9. With 355
classes, a Dionis fit takes 1.2s (sigmoid) or 4.5s (platt, whose likelihood
fit uses a row subsample once the margin matrix passes 1e7 entries).
sefr_boost fits a SEFR model per boosting round, so on Dionis it still
overruns small budgets (18.1x at 2s, 3.8x at 10s).

Accuracy

SEFR does not beat the tree learners, and this PR does not claim it does.

Binary, test AUC at a 10s budget:

dataset	sefr	sefr_boost	lrl1	lgbm
car	0.941	0.947	0.941	1.000
adult	0.846	0.884	0.506	0.913
Amazon_employee_access	0.580	0.591	0.518	0.842
Albert	0.630	0.676	0.596	0.758
Airlines	0.596	0.619	0.578	0.721
poker	0.502	0.501	0.502	0.931

Multiclass, negative log loss at a 10s budget (higher is better):

dataset	sefr	sefr_boost	lrl1	lgbm
connect-4	-0.657	-0.606	-0.593	-0.384
dilbert	-1.181	-0.811	-0.238	-0.124
Dionis	-2.170	-2.204	-5.872	-2.083

SEFR beats lrl1 on 4 of 6 binary tasks, ties it on poker (both at chance) and car (0.94098 vs
0.94113), and loses to it on 2 of 3 multiclass tasks. On Dionis it comes within
0.09 of lgbm; elsewhere lgbm is clearly better.

sefr_boost beat the default pipeline on dilbert (2-10s) and connect-4 (1s). It
also scores higher on Dionis, but only while overrunning the budget (3.8x-18x),
so that is not a like-for-like win. Under roc_auc, "sefr" barely responds to
its search space, so treat it as a fixed-cost baseline. "sefr_boost" is the one
that improves with more budget.

That is why both are opt-in.

Tests

test/automl/test_model.py adds 19 unit tests, including checks of the fitted
model against the closed form of Eqs. 3-9 recomputed independently (binary, and
each one-vs-rest head for multiclass), plus sample_weight (its effect on
calibration, zero and invalid weights), class_weight as costs, the binary Platt
intercept against LogisticRegression, the multiclass calibration and its row
subsample, sparse/dense equivalence and rejection of negative sparse input, and
sklearn's check_estimator suite, which passes in full on scikit-learn 1.0.2,
1.5.2 and 1.9.1. sefr_boost passes the SEFR learner as base_estimator on
scikit-learn < 1.2.
test/automl/test_extra_models.py adds two integration tests alongside the
other opt-in estimators.

Disclosure

I am an author of the SEFR paper. Happy to drop this if maintainers would rather
not carry it.

Related issue number

Related to #1612

Checks

• I've used pre-commit to lint the changes in this PR (note the same in integrated in our CI checks).
• I've included any doc changes needed for https://microsoft.github.io/FLAML/. See https://microsoft.github.io/FLAML/docs/Contribute#documentation to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

[Hamidreza Keshavarz (hamidkm9)]

@microsoft-github-policy-service agree

@microsoft-github-policy-service agree

[Li Jiang (thinkall)]

Overall review of the complete PR: changes are required.

1. flaml/automl/contrib/sefr.py:107,234: signed sparse inputs silently use max-abs scaling, violating SEFR's required [0,1] domain and reversing predictions. Remove unsafe modes or implement/reject with guaranteed non-negative transforms.
2. flaml/automl/contrib/sefr.py:137-158: sample weights affect the SEFR head but are ignored by probability calibration; zero-weight classes can also divide by zero. Validate weights and pass effective weights into calibration.
3. flaml/automl/contrib/sefr.py:314: AdaBoostClassifier(estimator=...) is incompatible with FLAML's supported scikit-learn 1.0/1.1 floor. Support base_estimator for older APIs or raise the minimum consistently.
4. flaml/automl/contrib/sefr.py:204: fitted-state, input, feature-count, sample-weight, NaN/inf, and sparse capability validation do not satisfy basic sklearn estimator contracts. Use sklearn validation helpers and accurate tags.
5. flaml/automl/contrib/sefr.py:157 and related docs: default Platt calibration is iterative and multiclass/scaler/calibrator state exceeds the advertised single-pass and n_features + 1 model-size guarantees. Correct the implementation default or the claims.

Posted by thinkall-agent-auto-reviewer

[Hamidreza Keshavarz (hamidkm9)]

Overall review of the complete PR: changes are required.

1. flaml/automl/contrib/sefr.py:107,234: signed sparse inputs silently use max-abs scaling, violating SEFR's required [0,1] domain and reversing predictions. Remove unsafe modes or implement/reject with guaranteed non-negative transforms.
2. flaml/automl/contrib/sefr.py:137-158: sample weights affect the SEFR head but are ignored by probability calibration; zero-weight classes can also divide by zero. Validate weights and pass effective weights into calibration.
3. flaml/automl/contrib/sefr.py:314: AdaBoostClassifier(estimator=...) is incompatible with FLAML's supported scikit-learn 1.0/1.1 floor. Support base_estimator for older APIs or raise the minimum consistently.
4. flaml/automl/contrib/sefr.py:204: fitted-state, input, feature-count, sample-weight, NaN/inf, and sparse capability validation do not satisfy basic sklearn estimator contracts. Use sklearn validation helpers and accurate tags.
5. flaml/automl/contrib/sefr.py:157 and related docs: default Platt calibration is iterative and multiclass/scaler/calibrator state exceeds the advertised single-pass and n_features + 1 model-size guarantees. Correct the implementation default or the claims.

Posted by thinkall-agent-auto-reviewer

Thanks Li Jiang (@thinkall) for the review. All five points are addressed in the new commits:

1. Signed sparse input. maxabs is removed. Dense input is min-max scaled and clipped to [0, 1] at predict time. Sparse input is scaled by column max and rejected with a clear error if it has negative entries, since it can't be shifted without densifying. scaling="none" rejects negative features. scaling is no longer in the search space, since only minmax is safe for arbitrary input.
2. Sample weights. sample_weight is validated for shape, finite values and non-negativity, and classes with zero total weight raise instead of dividing by zero. The effective weights, including class_weight, now reach the calibration and the scaler's feature range. A zero weight gives the same fit as dropping the sample, and there's a test for that.
3. scikit-learn 1.0/1.1. AdaBoost gets base_estimator on scikit-learn < 1.2 and estimator otherwise. Verified on 1.0.2.
4. sklearn contracts. It now uses validate_data/_validate_data, check_is_fitted and check_classification_targets, accepts dict class_weight, and declares sparse and positive-only tags. A new test runs check_estimator, which passes on 1.0.2, 1.5.2 and 1.9.1 except check_class_weight_classifiers. That one is a documented expected failure: SEFR's class weights cancel in the per-class means and only move the Eq. 9 threshold, so it can't reach the 87% majority the check expects.
5. Calibration and claims. Binary margins map to sigmoid(a·m). Multiclass margins map to softmax(a·m + b_k), with one shared scale and a bias per class, and decision_function returns those logits. So predict, decision_function and predict_proba always agree, which wasn't guaranteed before. The default sigmoid is closed form (a from the margin spread, b_k the log class prior). platt fits a and b_k by maximum likelihood and is documented as iterative. The docstrings, docs and PR description no longer claim single-pass or n_features + 1 for the full model.
   I re-ran the budget sweep on the current code and updated the tables in the description. Binary results are unchanged. On Dionis (355 classes), sefr log loss improved from 3.07 to 2.16, within 0.08 of lgbm, and its worst overrun dropped from 18.5x to 8.6x.

[Li Jiang (thinkall)]

Overall review of the complete current PR: changes are still required.

1. flaml/automl/contrib/sefr.py:291: binary Platt calibration disables the intercept, so it cannot learn imbalanced or weighted class priors. Fit and retain a binary intercept and derive prediction consistently from the calibrated score.
2. flaml/automl/contrib/sefr.py:305: calibration subsampling can preserve a zero-weight class row while omitting every positive-weight row for that class. Preserve positive-weight coverage and revalidate effective class totals after subsampling.
3. flaml/automl/contrib/sefr.py:322: advertised class_weight does not materially affect the decision rule and the standard sklearn contract check is suppressed. Implement it correctly or remove it from the public API/search space.
4. flaml/automl/contrib/sefr.py:179: sparse support is not declared through legacy _more_tags() on supported sklearn 1.0-1.5, so meta-estimators treat SEFR as dense-only.
5. flaml/automl/contrib/sefr.py:286 and documentation: multiclass fitting materializes multiple n_samples × n_classes arrays and performs a full pass per class, contradicting the constant-pass claim and creating multi-gigabyte allocations on documented datasets. Subsample/stream before materialization and document actual complexity.

Posted by thinkall-agent-auto-reviewer

[Hamidreza Keshavarz (hamidkm9)]

Overall review of the complete current PR: changes are still required.

1. flaml/automl/contrib/sefr.py:291: binary Platt calibration disables the intercept, so it cannot learn imbalanced or weighted class priors. Fit and retain a binary intercept and derive prediction consistently from the calibrated score.
2. flaml/automl/contrib/sefr.py:305: calibration subsampling can preserve a zero-weight class row while omitting every positive-weight row for that class. Preserve positive-weight coverage and revalidate effective class totals after subsampling.
3. flaml/automl/contrib/sefr.py:322: advertised class_weight does not materially affect the decision rule and the standard sklearn contract check is suppressed. Implement it correctly or remove it from the public API/search space.
4. flaml/automl/contrib/sefr.py:179: sparse support is not declared through legacy _more_tags() on supported sklearn 1.0-1.5, so meta-estimators treat SEFR as dense-only.
5. flaml/automl/contrib/sefr.py:286 and documentation: multiclass fitting materializes multiple n_samples × n_classes arrays and performs a full pass per class, contradicting the constant-pass claim and creating multi-gigabyte allocations on documented datasets. Subsample/stream before materialization and document actual complexity.

Posted by thinkall-agent-auto-reviewer

Thanks Li Jiang (@thinkall). All five points are addressed in 7fdd46f:

1. Binary Platt intercept. Binary platt now fits a slope and an intercept by weighted maximum likelihood, and predict follows the calibrated score. A test checks it against LogisticRegression. A small fixed penalty on the slope only keeps separable data finite and doesn't depend on the scale of the weights (AdaBoost normalizes them).
2. Subsample coverage. The calibration subsample keeps a positive-weight row of every class and re-checks class totals on the subsample. A test covers the case where every row on the subsample grid, and the first row of each class, have zero weight.
3. class_weight. It is now implemented as misclassification costs on the calibrated decision. The closed-form calibration adds log(c) to each class's logit, and platt weights the likelihood by class cost. It no longer cancels in the class means. check_class_weight_classifiers now passes, so check_estimator runs with no exemptions on scikit-learn 1.0.2, 1.5.2 and 1.9.1.
4. Legacy tags. _more_tags() declares X_types: ["2darray", "sparse"] for scikit-learn < 1.6.
5. Memory and passes. Per-class sums come from one sparse one-hot matrix product. Every one-vs-rest head and its Eq. 9 bias follow in closed form, and score spreads are accumulated in row chunks. Fitting takes three passes over the data whatever the number of classes, and no n_samples × n_classes array exists except the Platt block, capped at 1e7 entries. It reproduces the previous model to about 1e-14. On Dionis (312k × 60, 355 classes), a fit went from 13.3s and 4.27 GB peak to 1.2s and 0.42 GB (sigmoid), and from 17.7s to 4.5s and 0.64 GB (platt). The docstrings and docs now describe this cost.
   I re-ran the budget sweep and updated the tables in the description. sefr now stays within 1.7x of the budget on all nine datasets (Dionis went from 8.6x to 1.7x).

[Li Jiang (thinkall)]

Overall review of the complete current PR: changes are still required.

1. flaml/automl/contrib/sefr.py:304: a model fitted on dense data learns a scaler offset, but sparse prediction input skips that offset. Identical values can produce different scores and labels solely from ndarray versus CSR representation. Apply the fitted transformation consistently or reject unsupported cross-format input.
2. flaml/automl/contrib/sefr.py:239: eps and threshold_shift are not validated. Zero/NaN values can fit successfully with non-finite coefficients or probabilities. Require finite positive eps and finite threshold_shift.

Posted by thinkall-agent-auto-reviewer

[Hamidreza Keshavarz (hamidkm9)]

Thanks Li Jiang (@thinkall). Both points are addressed in 90bc8be:

1. Dense fit, sparse predict. A model fitted on dense data with a nonzero feature minimum now applies its fitted offset to sparse input too, by densifying it, since subtracting the offset would densify it anyway. Dense and sparse input therefore always get identical scores. A test checks this for zero and nonzero minima, including values outside the training range.
2. Parameter validation. eps must be a finite positive number and threshold_shift a finite number. Anything else raises a ValueError. A test covers zero, negative, NaN, infinite, string, None and boolean values.

Is there anything else blocking this from being merged?

[Li Jiang (thinkall)]

Overall review of the complete current PR: approved.

The final two blockers are resolved. A dense-fitted SEFR model now applies its learned offset to sparse prediction input before scaling, and the new regression tests compare identical dense/CSR probes with zero and nonzero fitted minima. eps now requires a finite positive real value and threshold_shift a finite real value, with tests for invalid and valid NumPy-scalar inputs. The previously reviewed calibration, weighting, sklearn-version, tag, and multiclass-resource fixes remain unchanged. No blocking code issues found. Build-matrix CI is still running on this head.

Posted by thinkall-agent-auto-reviewer

[Hamidreza Keshavarz (hamidkm9)]

Li Jiang (@thinkall) The only failing check is build (ubuntu-latest, 3.12), which timed out in test/tune/ (unrelated to this PR). The same tune tests pass on the other 11 jobs, and all SEFR tests passed in that job before the timeout. Could you re-run the failed job?

[Copilot]

🟡 Changes recommended

Boosting-wrapper cloning fails, and calibration subsampling can distort probabilities and exceed its memory cap.

3 open issues

• Preserve existing estimator parameters during wrapper reconstruction · New
• Enforce calibration size cap after class coverage sampling · New
• Avoid biased fixed-stride sampling for calibration · New

What changed in this PR

Adds SEFR and SEFR-based AdaBoost as opt-in classification options for low-budget AutoML, without changing the default estimator list.

Changes:

• Implements SEFR with scaling, weighting, and probability calibration.
• Registers both estimators and their tuning spaces.
• Adds unit tests, integration tests, and documentation.

File	Description
website/​docs/​Use-Cases/​Task-Oriented-AutoML.md	Documents estimators and hyperparameters.
test/​automl/​test_model.py	Tests SEFR mathematics, calibration, and compatibility.
test/​automl/​test_extra_models.py	Adds AutoML integration coverage.
flaml/​automl/​task/​generic_task.py	Registers both opt-in estimators.
flaml/​automl/​contrib/​sefr.py	Implements classifiers and tuning wrappers.
flaml/​automl/​contrib/​__init__.py	Exports the new wrappers.

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

The boosting wrapper replaces its supplied base estimator during reconstruction, breaking sklearn cloning and AutoML stacking.

4 open issues

• Preserve supplied base estimator during wrapper reconstruction · New
• Preserve existing estimator parameters during wrapper reconstruction
• Avoid biased fixed-stride sampling for calibration
• Enforce calibration size cap after class coverage sampling

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

Valid sample weights can produce NaN predictions, boosted feature-importance access raises an exception, and calibration unnecessarily copies the feature matrix.

1 open issue

• Avoid catastrophic cancellation in complementary weighted sums · New

2 previously missed issues

In code that hasn't changed since last review

Avoid copying full feature matrix during Platt calibration

flaml/​automl/​contrib/​sefr.py:418

When Platt calibration retains all rows, rows is still an integer index array, so X[rows] copies the entire feature matrix before computing margins. For 100,000 rows and 1,000 float64 features, this adds an unnecessary 800 MB allocation per fit, repeated in each boosting round. The margin-entry cap does not bound this feature copy. Reuse X directly when all rows are retained.

Handle missing feature importances in SEFR AdaBoost wrapper

flaml/​automl/​contrib/​sefr.py:569

Reading automl.feature_importances_ after fitting sefr_boost raises ValueError. AdaBoost requires its base learners to expose feature_importances_, but SEFRClassifier only exposes coef_. The inherited FLAML property's hasattr probe propagates AdaBoost's exception. Override the wrapper property to return None when base-learner importances are unavailable, consistent with the interface in flaml/automl/model.py:239–251.

4 resolved since last review

• Preserve supplied base estimator during wrapper reconstruction
• Preserve existing estimator parameters during wrapper reconstruction
• Avoid biased fixed-stride sampling for calibration
• Enforce calibration size cap after class coverage sampling

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Custom numerical calibration, weighted subsampling, and cross-version AdaBoost behavior warrant final human validation.

No open issues 🎉

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Numerical calibration, weighted subsampling, and version-dependent boosting behavior warrant final human validation across supported scikit-learn versions.

No open issues 🎉

_{Review effort: Balanced}

[Li Jiang (thinkall)]

Re-reviewed commits a87cc59453facc47419af33b272b3e905e5a18d6 -> 8f2c1e3dc2a42e46ed471f0898c8b16cae1c472d. The only delta is main's website/yarn.lock update of follow-redirects 1.16.0 -> 1.16.1. One explicitly head-bound review-code Low execution (actual runtime gpt-6.1-sol) found no new issue in this identical inherited patch. Controller Git comparisons verify all non-lockfile tree entries are unchanged. This is a scoped merge-delta review, not a fresh clean signoff for unchanged feature code.

Version/URL/caret/SRI shape checks pass. No independent local registry/tarball verification, package smoke test or website rebuild is claimed. Previous feature findings and runtime validation limits remain preserved.

At the 2026-10-26T12:13:44.943946+00:00 gate snapshot, actual current-head workflows were: Build in_progress/no conclusion; Code formatting completed/success; docs completed/success. Copilot has Balanced review without an Approval recommended verdict; 0 review thread(s) remain unresolved. Gate approval is not successful CI completion, and no new PR approval is being submitted while required gates remain unmet.

Posted by thinkall-agent-auto-reviewer

[Copilot]

🟡 Changes recommended

The auxiliary-memory bound omits row-sized allocations, understating memory requirements for large or sparse fits and needing correction.

1 open issue

• Include per-sample allocations in the auxiliary-memory bound · New

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Human assessment is needed to evaluate numerical instability with extremely small sample weights and potentially excessive memory use during calibration sampling.

No open issues 🎉

1 resolved since last review

• Include per-sample allocations in the auxiliary-memory bound

_{Review effort: Balanced}

[Li Jiang (thinkall)]

Reviewed new commits from 8f2c1e3dc2a42e46ed471f0898c8b16cae1c472d to 3e4ad873fa42663bc2715971f36bd371c46c6fdc. The main-branch advance is inherited across this update; the driver proves the complete resulting tree is the previously reviewed feature tree plus the same upstream patch, with no changed contributor feature blobs or function ASTs outside that integration. The late module-docstring correction is separately reviewed with actual CLI-default gpt-6.1-sol Low (clean); executable AST excluding only that docstring is identical.

One shared review-code Medium package used six independent isolated grok-4.7/xhigh lenses. All six original outcomes are issues_found and all canonical responses are committed on exact owning PR 1642 source aaa3b672462534d3e28dfbd3e88ce2e1c64a21ef, with explicit tree/content bindings to the updated heads. This is not six new whole-feature clearances or a blanket endorsement of unchanged feature code.

P2 inherited-upstream follow-up: real seasonal-naive HPO with 60,005 minute rows, period=12001, sample=True, min_sample_size=10000, max_iter=3, time_budget=15 chooses season 12,001 for a 10,000-row first trial and aborts with ValueError on current main and the owning updated head. Identical real-search settings succeed on exact prior main on both native stacks. Bound the season against the minimum actual sampled training window or avoid undersized sampling for that configuration. This issue is already in current main and inherited, not a new contributor-specific feature edit; it is preserved as an upstream follow-up rather than duplicated as an unrelated inline request. The shared review therefore does not grant new full-source approval. max_iter=1 bypasses search and refits full data, so those initial successful diagnostics were not used as HPO clearance.

Selected integration validation passed 592 executions of 296 distinct cases on native Windows Python 3.10/pandas 2.3.3/NumPy 1.26.4 and Python 3.13/pandas 3.0.5/NumPy 2.3.3. That includes actual classifier and naive/seasonal-average fits plus controlled accessor/facade tests. Additional native controls reject unproven zero-lag/season-1/NaT-lookup/sparse-exog allegations; actual SARIMAX sparse prediction selects the requested rows. No native Spark/TFT training or complete optional/minimum-version matrix is claimed. Passing selected tests does not override the independently reproduced P2.

At 2026-10-31T20:28:04.925691+00:00, actual current-head pull_request workflows were: Code formatting completed/success (attempt 2); Build in_progress/None (attempt 2); docs completed/success (attempt 2). 0 review thread(s) remain unresolved. No explicit current-head Balanced Copilot approval recommendation is claimed. No new PR approval is submitted while source or external gates remain unmet.

Posted by thinkall-agent-auto-reviewer

[Copilot]

🟢 Approval recommended

No active review issues are recorded for the opt-in SEFR and SEFR-based AdaBoost additions.

No open issues 🎉

_{Review effort: Balanced}

[Li Jiang (thinkall)]

Review cycle 20261102-2221; exact source c33561fd4e339f804b724df802c42a590d24d67e.

Reviewed the new inherited Joi-only merge delta with review-code (Low, actual gpt-6.1-sol; local evidence commit e938ad78f120a83bc8825d27b8c114205b089ab9). Complete-tree reconstruction confirms the contributor's feature blobs are unchanged. This clean lock-delta result is not a new whole-feature clean vote.

The previously reproduced inherited-main seasonal-naive HPO P2 remains: 60,005 minute observations, period 12,001, sampling/minimum sample 10,000, max_iter=3 and time_budget=15 select an invalid season for the first sampled trial and abort AutoML.fit; prior-main controls pass on both native stacks. The Joi-only main advance changes none of that Python code. No unchanged Python/native/Spark/TFT suite was rerun for this lock-only update.

Fresh gates: 0 unresolved review thread(s); explicit current-head Balanced approval recommendation; 11 non-success applicable check context(s). Actual workflow holds: Build:in_progress/None. Workflow permission approval is not validation success. No PR approval or foreign-thread resolution was performed.

Posted by thinkall-agent-auto-reviewer

[Copilot]

🟡 Changes recommended

Zero-weight rows can still cause overflow and NaN calibration, and zero-variance heads apply an arbitrary threshold offset instead of leaving predictions unchanged.

2 open issues

• Exclude zero-weight rows from scoring and calibration · New
• Use actual zero score standard deviation for threshold shifts · New

_{Review effort: Balanced}

[Li Jiang (thinkall)]

Review cycle 20261106-1600; exact source 0a205d47e71f7d389c095b41ee603a82bd993ba7.

The branch was updated with a fresh permission/conflict/head guard. Complete-tree reconstruction matches the exact previous feature tree merged with current main 46bce0ce989aae0271e68e5daac5b8a9b0522b87; no unrelated merge edits were found.

Independent review is incomplete. The shared merge delta is planned Low under review-code, but its authenticated CLI prerequisite remains unresolved. No reviewer model ran and no new clean skill artifact was produced. Earlier feature findings and scope qualifications are retained, not converted into a new whole-feature clean vote.

Bounded shared-main validation: the Spark-backend/R2 tests passed 15 cases on each native Windows Python 3.10 and 3.13 stack. The same external test files against the immediate previous main produced six expected failures and nine passing controls per stack. Spark tests use a controlled backend, not an actual Spark cluster; these checks do not cover the whole feature or optional workload matrix.

Existing inherited-main HPO hold remains. Real AutoML.fit search with 60,005 minute observations, period=12001, sample=True, min_sample_size=10000, max_iter=3 and time_budget=15 still selects season 12,001 for the first 10,000-row trial and aborts with SeasonalNaive needs at least season=12001 training observations, got 10000. This was reproduced on both native stacks for new main and its immediate predecessor; the newly merged changes did not repair the older issue. Historical pre-introduction passing controls remain separate. max_iter=1 is not search/sampling proof.

Fresh gates: 2 unresolved thread(s); explicit current-head Balanced approval recommendation: no; 11 non-success applicable context(s). Actual current-head workflow holds: Build:in_progress/None. Workflow permission approval is not validation success. No PR approval, foreign-thread mutation or review dismissal was performed.

Posted by thinkall-agent-auto-reviewer

[Copilot]

🔵 Needs a closer look

The custom likelihood calibration and weighted sampler require numerical review and runtime validation of memory/budget bounds and AdaBoost compatibility across supported scikit-learn versions.

No open issues 🎉

2 resolved since last review

• Use actual zero score standard deviation for threshold shifts
• Exclude zero-weight rows from scoring and calibration

_{Review effort: Balanced}

[Li Jiang (thinkall)]

Review cycle 20261106-2049; exact source d157a5240f261bef366b9695ea22db91f5e3f4e2.

Inspected the new author delta from 0a205d47e71f7d389c095b41ee603a82bd993ba7. Independent review-code review is planned Low, but remains blocked at its unresolved authenticated CLI prerequisite. No reviewer model ran or new canonical clean artifact was generated. Bounded driver checks below are not a whole-feature clean signoff; original findings and review-scope qualifications are preserved.

Corrective evidence: positive-weight filtering now occurs before scaling, scoring and calibration. Each native Windows Python 3.10/3.13 stack passed 21 scoped cases, including finite 1e160 zero-weight extremes, dense/CSR input, binary/multiclass labels, minmax/no scaling, sigmoid/Platt calibration, input preservation and existing weight-validation controls. The same tests against the exact pre-fix source produced eight expected failures and 13 passing controls per stack. This verifies the bounded zero-weight correction, not numerical memory/budget bounds or the full AdaBoost/version matrix. Copilot's current Balanced review has no open issues but says 'Needs a closer look', not an explicit approval recommendation.

The older inherited seasonal-naive HPO P2 also remains on this exact author head: 60,005 minute observations, period=12001, sample=True, min_sample_size=10000, max_iter=3, time_budget=15 still abort with SeasonalNaive needs at least season=12001 training observations, got 10000. Both native stacks reproduced it. These are real search checks, not max_iter=1 refit proxies.

Fresh external gates: 0 unresolved thread(s); current-head Balanced approval recommendation: no; actual workflow holds: Build:completed/action_required; Code formatting:completed/action_required; docs:completed/action_required. No PR approval, ordinary issue comment, foreign-thread mutation or review dismissal was performed.

Posted by thinkall-agent-auto-reviewer