gravatar

JonasKnecht

JonasKnecht

Recently Published

naive_bayes_model_variations
finance_nlp_eda
feature_comments_expost
matched_cnn_v7
tired_eyes_regression
matched-cnn-training-0.0.3
matched-cnn-comparison-0.0.1
matched-sample-0.0.1
well_groomed_checks
well-groomed-orthogonal-checks
latent_predictor_wg_orth
Corrective-Gradient-WellGroomed
rearrest_risk_serious_crime
Tired-label correlation check
Todorov Projector Labels
Faces Draft 1 Figures -0.0.4
Faces_Draft1_Figures
Conviction Features
XgBoost Tire Kicking
draft_1_slides_figures
Faces_Presentation_Figures
ml_face_fusion_structured
comparing_jf_jb
Clean Cuttness V2
election_data_prep
clean-cut-regressions
dex_age_cnn_plots
latent_noise_permutation
noise_tensor_permutation
stylegan2_malesubset_metrics
gan_editing_noise
eth_presentation_tables
arrest_table01_mturk_sbs
election_mturk_quality_control
mturk_sbs_high_detail
pytorch_generator_distributions
mturk_sbs_plots
arrest_projector_regressions
election_table01_mturk_sbs
arrest_cnn_comparison
faces_arrest_log
arrest_table01_edits
election_mturk_plots
election_project_log
election_table01_config4_5
election_table01_config2
election_project_log
election_cnn_table01
table01
robustness_checks_summary
baselin_regressions_summary
increased_detail_regressions
Publish Document
Faces Project Log File
Bivariate_Regressions
Regression_NonLinearity_Update
Faces_Regressions_Update
Updated regressions with MTurk labels.
Election_CNN_Regressions01
Regression and AUC summary from the first batch of election models.
XgBoost_Risk_Regression
Regression output for the inclusion of an XgBoost risk predictor together with the conv-net and covariate inputs.
GAN_Survey_Training
This markdown contains the initial training set for the GAN survey for image quality.
styleGAN2_ssim_correlation
This markdown contains a look at the potential correlation between SSIM, MSE, and the predicted outcomes from our CNN model.
StyleGAN2_Run2_PRMetrics
StylegGAN2_Run2_Metrics
StyleGAN2_TrainingRun1_Summary
A summary of potential next steps taken to improve the training of StyleGAN2 on a rather homogeneous sample.