Skill UI

A comprehensive guide for researchers and reviewers designing or auditing Automated Software Engineering (ASE) evaluations. It outlines the necessary rigor, including the use of real subject systems, establishing fair and runnable tool baselines, performing targeted ablations (especially for LLMs), defining explicit oracles, and ensuring full data and model provenance to guarantee reproducible and scientifically sound results.

This guide provides expert principles for designing and auditing system-level experiments in computer architecture (e.g., for ASPLOS-level conferences). It details the proper usage of various instruments—from real silicon and FPGAs to cycle-level simulators and analytical models—ensuring that every claim is matched to a valid, stated error model. It emphasizes workload selection, establishing strong baselines, performing rigorous ablation studies, and honestly reporting measurement noise and limitations to maintain scientific credibility.

This guide models the complex and rigorous two-stage filtering process for submissions to the CAV conference. It details the partial double-anonymity, the core evaluation criteria (significance, soundness, novelty), and provides strategic advice on how authors should structure their paper to withstand early rejection, and how to use the rebuttal phase effectively.

A comprehensive guide detailing best practices for designing and auditing empirical evaluations of Large Language Models (LLMs). It addresses critical pitfalls such as data contamination, baseline unfairness, model version drift, and decoding parameter sensitivity, ensuring that published results are robust, reproducible, and scientifically sound for academic publication.

This comprehensive guide provides expert protocols for writing the 'Related Work' section in Machine Learning papers. It teaches authors how to maintain 'versioned honesty' by meticulously triaging the literature into distinct lanes (e.g., method lineage, model artifacts, evaluation infrastructure). It covers advanced topics like managing concurrent work, citing arXiv status accurately, and avoiding self-citation violations, ensuring the paper is positioned rigorously within the rapidly evolving academic field.

A comprehensive guide for authors submitting research to CVPR, detailing best practices for packaging code, models, and datasets. It covers requirements for review-time reproducibility (sealed-box rules) and long-term public release (licensing, provenance, and version pinning), ensuring that the work is runnable by any skeptical third party.

A detailed model of the ACM/IEEE Design Automation Conference (DAC) research manuscript review process. This guide explains the double-blind, TPC-driven, and single-shot nature of the evaluation, emphasizing the critical balance between genuine novelty and measured Quality-of-Results (QoR). It outlines the high standards for evidence fairness, benchmark credibility, and ablation studies required for acceptance, differentiating DAC's approach from architecture and software venues.

A comprehensive guide for preparing research artifacts (code, data, prompts, models) for EACL submissions. It details the two stages: an anonymized supplement for the initial review (ARR) and a fully licensed, public release for post-acceptance. Emphasis is placed on reproducibility, responsible NLP practices, and proper documentation.

A comprehensive guide for structuring the empirical evidence section of top-tier NLP conference papers (e.g., EACL). It mandates rigorous controls, including fair baselines (tuned models, explicit LLM prompts), statistical significance testing (variance, CI), contamination checks, and detailed human evaluation protocols. Use this to elevate your research from mere assertion to scientifically audited proof.

This guide provides a detailed strategy for managing and releasing research artifacts (code, models, data) associated with a vision paper submitted to ECCV. It covers two critical phases: the sealed, anonymous archival submission during review, and the structured public release runway post-acceptance, ensuring reproducibility and adherence to conference policies.

A comprehensive guide detailing advanced econometric methods for establishing causal relationships in policy evaluations. It covers techniques such as Difference-in-Differences (DID), Instrumental Variables (IV), Regression Discontinuity Design (RDD), and structural modeling. The core focus is bridging the gap between econometric rigor and policy legibility, ensuring that the causal mechanism is explicit in the main text while maintaining formal defense in the appendix.

A comprehensive guide detailing the multi-cycle, rolling model for EDBT research submissions. It covers the entire review pipeline, from initial program committee evaluation and author-feedback exchange, through the accept/revise/reject decision, and critically, the in-cycle revise-and-resubmit stage. The guide also contrasts EDBT's open-access, three-cycle cadence with other top-tier database conferences (e.g., SIGMOD, VLDB).