Loading CVE JSON files... 1681it [00:01, 1365.79it/s] Loaded 1681 CVEs Generating embeddings... Model loaded from: ./domain_adapted_model Batches: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 53/53 [00:01<00:00, 43.66it/s] Embeddings shape: (1681, 384), sample: [ 0.09512987 0.10447408 0.10577734 0.03425939 -0.02108146]... Generating query embeddings... Model loaded from: ./domain_adapted_model Batches: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 141.89it/s] Embeddings shape: (3, 384), sample: [-0.0122127 0.05706869 -0.01711867 -0.07662857 -0.11884194]... ====================================================================== DIFFUSION PARAMETER SWEEP ====================================================================== ====================================================================== Building BASELINE standard index... ====================================================================== [pyarrowspace] Convert pyarray2 and Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] Building from rows [pyarrowspace] built ArrowSpace: nitems=1681, nfeatures=384, lambdas_len=1681 Standard build time: 161.86s [pyarrowspace] search: qlen=384, lambda_q=0.567479 [pyarrowspace] search: qlen=384, lambda_q=0.608688 [pyarrowspace] search: qlen=384, lambda_q=0.178207 ====================================================================== Testing: η=0.05, steps=2 ====================================================================== Building energy index: η=0.05, steps=2, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 157.72s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2000 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 1.0000 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.4500 Recall@10: 0.5000 Recall@20: 0.5000 NDCG@10: 0.9360 Aggregated metrics: Avg MRR: 0.6667 Avg MAP: 0.2167 Avg NDCG@10: 0.6453 Avg Recall@10: 0.2333 ====================================================================== Testing: η=0.05, steps=4 ====================================================================== Building energy index: η=0.05, steps=4, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 151.07s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3417 Recall@10: 0.4000 Recall@20: 0.4000 NDCG@10: 0.8751 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2000 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.9509 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1000 Recall@10: 0.1000 Recall@20: 0.1000 NDCG@10: 1.0000 Aggregated metrics: Avg MRR: 1.0000 Avg MAP: 0.2139 Avg NDCG@10: 0.9420 Avg Recall@10: 0.2333 ====================================================================== Testing: η=0.05, steps=6 ====================================================================== Building energy index: η=0.05, steps=6, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 151.90s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3235 Recall@10: 0.3000 Recall@20: 0.4000 NDCG@10: 1.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1000 Recall@10: 0.1000 Recall@20: 0.1000 NDCG@10: 1.0000 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3444 Recall@10: 0.4000 Recall@20: 0.4000 NDCG@10: 0.9752 Aggregated metrics: Avg MRR: 1.0000 Avg MAP: 0.2560 Avg NDCG@10: 0.9917 Avg Recall@10: 0.2667 ====================================================================== Testing: η=0.05, steps=8 ====================================================================== Building energy index: η=0.05, steps=8, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 151.06s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.5000 MAP@20: 0.1933 Recall@10: 0.3000 Recall@20: 0.4000 NDCG@10: 0.6622 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1667 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.9304 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.4514 Recall@10: 0.5000 Recall@20: 0.5000 NDCG@10: 0.9553 Aggregated metrics: Avg MRR: 0.8333 Avg MAP: 0.2705 Avg NDCG@10: 0.8493 Avg Recall@10: 0.3333 ====================================================================== Testing: η=0.1, steps=2 ====================================================================== Building energy index: η=0.1, steps=2, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 155.04s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.2500 MAP@20: 0.0798 Recall@10: 0.2000 Recall@20: 0.3000 NDCG@10: 0.4662 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2972 Recall@10: 0.3000 Recall@20: 0.4000 NDCG@10: 0.9241 Aggregated metrics: Avg MRR: 0.4167 Avg MAP: 0.1257 Avg NDCG@10: 0.4634 Avg Recall@10: 0.1667 ====================================================================== Testing: η=0.1, steps=4 ====================================================================== Building energy index: η=0.1, steps=4, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 153.29s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3800 Recall@10: 0.4000 Recall@20: 0.4000 NDCG@10: 0.9585 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1667 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.9639 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2885 Recall@10: 0.3000 Recall@20: 0.5000 NDCG@10: 0.8235 Aggregated metrics: Avg MRR: 1.0000 Avg MAP: 0.2784 Avg NDCG@10: 0.9153 Avg Recall@10: 0.3000 ====================================================================== Testing: η=0.1, steps=6 ====================================================================== Building energy index: η=0.1, steps=6, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 154.50s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1708 Recall@10: 0.3000 Recall@20: 0.3000 NDCG@10: 0.8274 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.5000 Recall@10: 0.5000 Recall@20: 0.5000 NDCG@10: 0.9518 Aggregated metrics: Avg MRR: 0.6667 Avg MAP: 0.2236 Avg NDCG@10: 0.5931 Avg Recall@10: 0.2667 ====================================================================== Testing: η=0.1, steps=8 ====================================================================== Building energy index: η=0.1, steps=8, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 155.22s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0588 MAP@20: 0.0170 Recall@10: 0.0000 Recall@20: 0.2000 NDCG@10: 0.0000 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Aggregated metrics: Avg MRR: 0.0196 Avg MAP: 0.0057 Avg NDCG@10: 0.0000 Avg Recall@10: 0.0000 ====================================================================== Testing: η=0.15, steps=2 ====================================================================== Building energy index: η=0.15, steps=2, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 154.88s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.1429 MAP@20: 0.0143 Recall@10: 0.1000 Recall@20: 0.1000 NDCG@10: 0.3333 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2167 Recall@10: 0.2000 Recall@20: 0.3000 NDCG@10: 1.0000 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1000 Recall@10: 0.1000 Recall@20: 0.1000 NDCG@10: 1.0000 Aggregated metrics: Avg MRR: 0.7143 Avg MAP: 0.1103 Avg NDCG@10: 0.7778 Avg Recall@10: 0.1333 ====================================================================== Testing: η=0.15, steps=4 ====================================================================== Building energy index: η=0.15, steps=4, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 153.74s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2944 Recall@10: 0.4000 Recall@20: 0.4000 NDCG@10: 0.8642 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.5000 MAP@20: 0.0667 Recall@10: 0.1000 Recall@20: 0.2000 NDCG@10: 0.6309 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1182 Recall@10: 0.1000 Recall@20: 0.2000 NDCG@10: 1.0000 Aggregated metrics: Avg MRR: 0.8333 Avg MAP: 0.1598 Avg NDCG@10: 0.8317 Avg Recall@10: 0.2000 ====================================================================== Testing: η=0.15, steps=6 ====================================================================== Building energy index: η=0.15, steps=6, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 154.09s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3796 Recall@10: 0.5000 Recall@20: 0.5000 NDCG@10: 0.9217 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Aggregated metrics: Avg MRR: 0.3333 Avg MAP: 0.1265 Avg NDCG@10: 0.3072 Avg Recall@10: 0.1667 ====================================================================== Testing: η=0.15, steps=8 ====================================================================== Building energy index: η=0.15, steps=8, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 154.22s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0000 MAP@20: 0.0000 Recall@10: 0.0000 Recall@20: 0.0000 NDCG@10: 0.0000 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.5000 MAP@20: 0.1000 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.6399 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.5000 Recall@10: 0.5000 Recall@20: 0.5000 NDCG@10: 0.9720 Aggregated metrics: Avg MRR: 0.5000 Avg MAP: 0.2000 Avg NDCG@10: 0.5373 Avg Recall@10: 0.2333 ====================================================================== Testing: η=0.2, steps=2 ====================================================================== Building energy index: η=0.2, steps=2, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 154.90s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.5000 MAP@20: 0.1667 Recall@10: 0.3000 Recall@20: 0.3000 NDCG@10: 0.6604 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1286 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.9180 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3750 Recall@10: 0.4000 Recall@20: 0.5000 NDCG@10: 0.9560 Aggregated metrics: Avg MRR: 0.8333 Avg MAP: 0.2234 Avg NDCG@10: 0.8448 Avg Recall@10: 0.3000 ====================================================================== Testing: η=0.2, steps=4 ====================================================================== Building energy index: η=0.2, steps=4, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 156.03s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1400 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.5211 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.3333 MAP@20: 0.0476 Recall@10: 0.1000 Recall@20: 0.2000 NDCG@10: 0.5000 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.1000 Recall@10: 0.1000 Recall@20: 0.1000 NDCG@10: 1.0000 Aggregated metrics: Avg MRR: 0.7778 Avg MAP: 0.0959 Avg NDCG@10: 0.6737 Avg Recall@10: 0.1333 ====================================================================== Testing: η=0.2, steps=6 ====================================================================== Building energy index: η=0.2, steps=6, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 155.49s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2908 Recall@10: 0.3000 Recall@20: 0.4000 NDCG@10: 0.8974 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.5000 MAP@20: 0.0833 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 0.6194 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.4514 Recall@10: 0.5000 Recall@20: 0.5000 NDCG@10: 0.9630 Aggregated metrics: Avg MRR: 0.8333 Avg MAP: 0.2752 Avg NDCG@10: 0.8266 Avg Recall@10: 0.3333 ====================================================================== Testing: η=0.2, steps=8 ====================================================================== Building energy index: η=0.2, steps=8, optical_tokens=40 [pyarrowspace] build_energy: Converting pyarray2 to Vec [pyarrowspace] items shape: (1681, 384) [pyarrowspace] items[0][:5]: [1.141558438539505, 1.2536889910697937, 1.2693280577659607, 0.4111126810312271, -0.25297757238149643] [pyarrowspace] NaNs: 0, Infs: 0 [pyarrowspace] build_energy: optical_tokens=Some(40), w_λ=1.00, w_G=0.50, w_D=0.25 [pyarrowspace] build_energy: Starting energy pipeline [pyarrowspace] build_energy complete: nitems=1681, nfeatures=384, graph_nodes=39, lambdas_len=1681 Energy build time: 157.99s Query 1: authenticated arbitrary file read path traversal... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.3800 Recall@10: 0.4000 Recall@20: 0.4000 NDCG@10: 0.9330 Query 2: remote code execution in ERP web component... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 1.0000 MAP@20: 0.2000 Recall@10: 0.2000 Recall@20: 0.2000 NDCG@10: 1.0000 Query 3: SQL injection in login endpoint... [pyarrowspace] search_energy: qlen=384, k=20, w_λ=1.00, w_D=0.50 MRR: 0.0714 MAP@20: 0.0071 Recall@10: 0.0000 Recall@20: 0.1000 NDCG@10: 0.0000 Aggregated metrics: Avg MRR: 0.6905 Avg MAP: 0.1957 Avg NDCG@10: 0.6443 Avg Recall@10: 0.2000 ====================================================================== BEST CONFIGURATIONS ====================================================================== Best MRR: η=0.05, steps=4 MRR: 1.0000 MAP: 0.2139 NDCG: 0.9420 Best MAP: η=0.1, steps=4 MRR: 1.0000 MAP: 0.2784 NDCG: 0.9153 Sweep heatmaps saved to diffusion_sweep_heatmaps.png Sweep results saved to diffusion_sweep_results.csv ====================================================================== ANALYSIS COMPLETE ====================================================================== Standard build time: 161.86s Energy builds tested: 16/16