Track: Track1; Team name: LangDiff; Model: HiD-Net

[Mullerio]

Checklist

• My pull request has a clear and explanatory title.
• My pull request passes the Linting test.
• I added appropriate unit tests and I made sure the code passes all unit tests. (refer to comment below)
• My PR follows PEP8 guidelines. (refer to comment below)
• My code is properly documented, using numpy docs conventions, and I made sure the documentation renders properly.
• I linked to issues and PRs that are relevant to this PR.

Description

Implementation of HiDNet, published at AAAI 2024 (https://arxiv.org/abs/2312.08616) based on the original implementation https://github.com/BUPT-GAMMA/HiD-Net.

The implementation uses the default g3 high-order gradient variant from the official HiD-Net implementation.

We add:

• topobench/nn/backbones/graph/hid_net.py
• configs/model/graph/hid_net.yaml
• test/nn/backbones/graph/test_hid_net.py
• a results.json generated by the evaluation notebook using MODEL_CONFIG = "graph/hid_net"

Tests

Fully pass ruff checks, ran with:
uv run --no-sync ruff check topobench/nn/backbones/graph/hid_net.py test/nn/backbones/graph/test_hid_net.py

Added passing unit tests for:

• forward-pass output shape and finite node embeddings;
• compatibility with the GNNWrapper call signature, including batch and edge_weight;
• weighted edge propagation;
• differentiability of the backbone and finite gradients for inputs and learnable parameters.