Upload AUDIT_TRAIL.md

AUDIT_TRAIL.md

@@ -0,0 +1,88 @@
+# RTB Bidding Algorithm Comparison — Audit Trail: Every Resource Read
+
+> Generated: 2026-05-05
+> This document lists every paper, codebase, dataset, and documentation page consulted.
+
+---
+
+## Papers Read (Methodology Sections)
+
+| # | Paper | arXiv | Section(s) Read | Key Extraction |
+|---|-------|-------|-----------------|----------------|
+| 1 | Wang et al. — Learning to Bid in Repeated First-Price Auctions with Budgets | 2304.13477 | Section 3 (full) | Algorithm 1: DualOGD, equations (2)-(5), regret bound |
+| 2 | Balseiro et al. — Dual Mirror Descent for Online Allocation | 2011.10124 | Section 3 (full) | Theorem 1, mirror descent dual update, stopping time analysis |
+| 3 | Feng et al. — Online Bidding for RoS Constrained Advertisers | 2208.13713 | Section 3 (full) | Two-variable dual, multiplicative + sub-gradient updates, bid rule |
+
+## Papers Reviewed (Abstracts + Details)
+
+| # | Paper | arXiv | Venue | Year |
+|---|-------|-------|-------|------|
+| 4 | Cai et al. — RTB by RL in Display Advertising | 1701.02490 | WSDM | 2017 |
+| 5 | Wang et al. — HiBid Hierarchical DRL Bidding | 2312.17503 | — | 2023 |
+| 6 | Mao et al. — FinalMLP | 2304.00902 | AAAI | 2023 |
+| 7 | Wang et al. — GDCN | 2311.04635 | CIKM | 2023 |
+| 8 | Wang et al. — DCN V2 | 2008.13535 | WWW | 2021 |
+| 9 | Zhu et al. — BARS-CTR Benchmark | 2009.05794 | CIKM | 2021 |
+| 10 | — Robust Budget Pacing with a Single Sample | 2302.02006 | — | 2023 |
+| 11 | — Multi-Channel Autobidding with Budget and ROI | 2302.01523 | — | 2023 |
+| 12 | — No-Regret in Repeated FPA with Budgets | 2205.14572 | — | 2022 |
+| 13 | — Dynamic Budget Throttling | 2207.04690 | — | 2022 |
+| 14 | — AIGB: Generative Auto-bidding | 2405.16141 | — | 2024 |
+| 15 | — Adaptive Bidding under Non-Stationarity | 2505.02796 | — | 2025 |
+| 16 | — Joint Value Estimation and Bidding | 2502.17292 | — | 2025 |
+| 17 | — Leveraging Hints: Adaptive Bidding | 2211.06358 | — | 2022 |
+| 18 | — No-Regret is not enough! | 2405.06575 | — | 2024 |
+| 19 | — Coordinated Dynamic Bidding in Second-Price | 2306.07709 | — | 2023 |
+| 20 | — FCN: Fusing Cross Network | 2407.13349 | — | 2024 |
+
+## Citation Graphs Traced
+
+| # | Source Paper | Direction | Key Findings |
+|---|-------------|-----------|-------------|
+| 21 | Wang et al. 2304.13477 | Both (20 refs, 20 citations) | Trace to Balseiro 2020, contextual extension 2603.07207 |
+
+## Datasets Inspected
+
+| # | Dataset | Tool | Findings |
+|---|---------|------|----------|
+| 22 | reczoo/Criteo_x4 | hf_inspect_dataset | Verified: 45.8M rows, 13 dense + 26 cat, Label column, 5.6GB |
+| 23 | reczoo/Avazu_x4 | hf_inspect_dataset | Verified: 40.4M rows, 22 fields, click column, 1.8GB |
+| 24 | iPinYou (search) | hub_repo_search | NOT on HF Hub — external download only |
+
+## Codebases Identified
+
+| # | Repo | URL | Stars |
+|---|------|-----|-------|
+| 25 | han-cai/rlb-dp | https://github.com/han-cai/rlb-dp | 188 |
+| 26 | reczoo/FuxiCTR | https://github.com/reczoo/FuxiCTR | Major |
+| 27 | shenweichen/DeepCTR-Torch | https://github.com/shenweichen/DeepCTR-Torch | Major |
+| 28 | Novartis/torchsurv | https://github.com/Novartis/torchsurv | 200 |
+| 29 | openbenchmark/BARS | https://github.com/openbenchmark/BARS | 389 |
+| 30 | dingmu365/budget_constrained_bidding | https://github.com/dingmu365/budget_constrained_bidding | — |
+| 31 | GinNie23/budget_constrained_bidding | https://github.com/GinNie23/budget_constrained_bidding | — |
+| 32 | venkatacrc/Budget_Constrained_Bidding | https://github.com/venkatacrc/Budget_Constrained_Bidding | — |
+| 33 | salmon1802/FCN | https://github.com/salmon1802/FCN | — |
+| 34 | hamverbot/rtb-bidding-comparison | https://huggingface.co/hamverbot/rtb-bidding-comparison | Your repo |
+
+## External Resources
+
+| # | Resource | URL |
+|---|----------|-----|
+| 35 | BARS CTR Criteo_x4 Leaderboard | https://openbenchmark.github.io/BARS/CTR/leaderboard/criteo_x4.html |
+| 36 | BARS CTR Avazu Leaderboard | https://openbenchmark.github.io/BARS/CTR/leaderboard/avazu_x4.html |
+| 37 | FuxiCTR Documentation | https://fuxictr.github.io/ |
+| 38 | DeepCTR-Torch Documentation | https://deepctr-torch.readthedocs.io/ |
+| 39 | TorchSurv Documentation | https://opensource.nibr.com/torchsurv/ |
+| 40 | TorchSurv PyPI | https://pypi.org/project/torchsurv/ |
+| 41 | fuxictr PyPI | https://pypi.org/project/fuxictr/ |
+| 42 | deepctr-torch PyPI | https://pypi.org/project/deepctr-torch/ |
+| 43 | iPinYou Dataset | data.computational-advertising.org |
+| 44 | HF Papers Dashboard | https://huggingface.co/papers |
+
+---
+
+**Total resources consulted: 44**
+- Papers: 20 (3 read in full methodology, 17 abstracts/details)
+- Datasets: 3 inspected
+- Codebases: 10 identified
+- External references: 10 (leaderboards, docs, PyPI)