RAIDCXM (Under Review)

Code for our double‑blind submission on securing Concept-based Models (CxMs) under adversarial threat while preserving human intervention effectiveness.

Motivation & Background

Concept-based Models insert a human-understandable concept layer between a pretrained backbone and the classifier, enabling explanation and human intervention (manual concept correction). Prior work emphasized concept prediction robustness but rarely measured the end-to-end robustness and how interventions interact with adversarial attacks. We evaluate both pre‑intervention and post‑intervention adversarial robustness.

Quick Start

1. Install dependencies:

pip install -r requirements.txt

2. Prepare datasets (see Dataset Preparation).
3. Create override YAML list (e.g., tmp.yaml).
4. Run an experiment:

python main.py --config tmp.yaml --task_id 0

5. Omit --task_id to run all entries.

Configuration System

Defaults reside in config.yaml. Override file must be a YAML list; each element merges with defaults (recursive) via utils.yaml_merge. The merged config and checkpoint are stored under checkpoints/; run logs go to logs/<run_name>/<run_id>/ with deterministic run_name + hashed run_id.

Running Experiments

Example cfg.yaml:

- model: CBM
  dataset: CUB
  train_mode: Std
# below is config of RAIDCXM
- model: VCBM
  dataset: CUB
  train_mode: JPGD
  res_dim: 32
  hsic_weight: 0.0001

Run the second entry:

python main.py --config cfg.yaml --task_id 1

Dataset Preparation

Expected structure under data/:

CUB_200_2011/
  images/ ...
  attributes/
  train.pkl / val.pkl / test.pkl
Animals_with_Attributes2/
  JPEGImages/ ...
  classes.txt
  predicates.txt
  predicate-matrix-binary.txt

Ensure data_path points to the parent directory. To add a dataset, implement a loader class under dataset/. For CUB, *.pkl files can be found in CBM's repository here.