Backcasting forecast errors: model collapsing to mean [P]

<!-- SC_OFF --><div class="md"><p>Hey everyone,</p> <p>I’ve been building an end-to-end Air Quality (PM2.5) forecasting pipeline for 4 countries (US, UK, India, Australia) using

<!-- SC_OFF --><div class="md"><p>I built a calibrated predictive model on point-in-time fundamentals and two results surprised me enough to write up. Sharing the project and the

<table> <tr><td> <a href="https://www.reddit.com/r/MachineLearning/comments/1uud3qj/obtaining_irregular_learning_curves_with/"> <img alt="Obtaining Irregular Learning Curves with

<!-- SC_OFF --><div class="md"><p>I am responsible for a research project that is supposed to train a GPT-like model (Transformer-decoder) with 100M, 250M and 500M model

<!-- SC_OFF --><div class="md"><p>Building a sports prediction model ,I found consistent edge when backtesting against closing lines.<br /> At inference time tho, I predict 12-24

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1x9n3qv/tempopfn_synthetic_pretraining_of_linear_rnns_for/"> <img alt="TempoPFN: Synthetic Pre-training of

<!-- SC_OFF --><div class="md"><p>I plotted the Expected Calibration Error (ECE) for an LLM (Gemini 2.5 Pro) forecasting 30 different real-world time-series targets over 38 days

<!-- SC_OFF --><div class="md"><p>&#x200b;</p> <p>## Abstract</p> <p>Post-training makes language models more decisive without necessarily making them more accurate — and we find