Regression on California Housing
0.089MSEMLP
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MLPFLOPs=1568, Param=8332025.06 | 0.089 | — | — | — | |
| MLP2025.06 | 0.089 | — | — | — | |
| NS-TDFLOPs=1568, Param=18572025.06 | 0.1005 | — | — | — | |
| NS-TD2025.06 | 0.1005 | — | — | — | |
| KANFLOPs=19544, Param=70562025.06 | 0.1017 | — | — | — | |
| KAN2025.06 | 0.1017 | — | — | — | |
| NS-SRFLOPs=3104, Param=16012025.06 | 0.1135 | — | — | — | |
| NS-SR2025.06 | 0.1135 | — | — | — | |
| Bw-SQbits=82026.03 | 0.169 | — | — | — | |
| Bw-SQbits=52026.03 | 0.171 | — | — | — | |
| Bw-SQbits=72026.03 | 0.171 | — | — | — | |
| Bw-SQbits=62026.03 | 0.173 | — | — | — | |
| LLT4bits=82026.03 | 0.184 | — | — | — | |
| Bw-SQbits=42026.03 | 0.185 | — | — | — | |
| FPbits=22026.03 | 0.186 | — | — | — | |
| FPbits=32026.03 | 0.186 | — | — | — | |
| FPbits=42026.03 | 0.186 | — | — | — | |
| FPbits=52026.03 | 0.186 | — | — | — | |
| FPbits=62026.03 | 0.186 | — | — | — | |
| FPbits=72026.03 | 0.186 | — | — | — | |
| FPbits=82026.03 | 0.186 | — | — | — | |
| LSQbits=82026.03 | 0.188 | — | — | — | |
| Pr-QQbits=62026.03 | 0.189 | — | — | — | |
| Pr-QQbits=82026.03 | 0.189 | — | — | — | |
| LLT9bits=82026.03 | 0.191 | — | — | — | |
| Pr-QQbits=72026.03 | 0.192 | — | — | — | |
| LLT4bits=72026.03 | 0.192 | — | — | — | |
| LSQbits=72026.03 | 0.192 | — | — | — | |
| LSQbits=62026.03 | 0.195 | — | — | — | |
| LLT9bits=72026.03 | 0.196 | — | — | — | |
| LLT4bits=62026.03 | 0.197 | — | — | — | |
| LLT9bits=62026.03 | 0.197 | — | — | — | |
| Pr-QQbits=52026.03 | 0.199 | — | — | — | |
| LLT4bits=52026.03 | 0.204 | — | — | — | |
| LSQbits=52026.03 | 0.207 | — | — | — | |
| LLT9bits=52026.03 | 0.208 | — | — | — | |
| Bw-SQbits=32026.03 | 0.21 | — | — | — | |
| Pr-MQbits=82026.03 | 0.21 | — | — | — | |
| Pr-QQbits=42026.03 | 0.213 | — | — | — | |
| LSQbits=42026.03 | 0.216 | — | — | — | |
| LLT4bits=42026.03 | 0.219 | — | — | — | |
| LLT9bits=42026.03 | 0.222 | — | — | — | |
| Pr-MQbits=72026.03 | 0.242 | — | — | — | |
| Offline Random Forest Regression (RFR)Training mode=Offline, Access=Unlimited access to dataset2026.02 | 0.253 | — | 80.7 | — | |
| Pr-MQbits=62026.03 | 0.254 | — | — | — | |
| LSQbits=32026.03 | 0.259 | — | — | — | |
| LLT9bits=32026.03 | 0.273 | — | — | — | |
| LLT4bits=32026.03 | 0.278 | — | — | — | |
| Pr-QQbits=32026.03 | 0.279 | — | — | — | |
| Pr-MQbits=52026.03 | 0.279 | — | — | — | |
| Bw-SQbits=22026.03 | 0.29 | — | — | — | |
| XGBoostBase Estimator=N/A, Active Est.=100, Comp. Ratio=0.0%, Latency / 1k (ms)=0.512026.06 | 0.3021 | — | 77.32 | — | |
| Pr-MQbits=42026.03 | 0.326 | — | — | — | |
| Pr-QQbits=22026.03 | 0.338 | — | — | — | |
| SCSBBase Estimator=DecisionTree, Active Est.=67, Comp. Ratio=33.0%, Latency / 1k (ms)=11.95 (1.25×)2026.06 | 0.3381 | — | 74.61 | — | |
| Lasso-Pruned BaggingBase Estimator=DecisionTree, Active Est.=99, Comp. Ratio=1.0%, Latency / 1k (ms)=17.362026.06 | 0.3394 | — | 74.52 | — | |
| Standard BaggingBase Estimator=DecisionTree, Active Est.=100, Comp. Ratio=0.0%, Latency / 1k (ms)=14.972026.06 | 0.3424 | — | 74.29 | — | |
| LSQbits=22026.03 | 0.356 | — | — | — | |
| Pr-MQbits=32026.03 | 0.376 | — | — | — | |
| Proposed (I)Training set size (S)=10002025.02 | 0.384 | — | — | 0.346 | |
| Random layer insertion (I)Training set size (S)=10002025.02 | 0.386 | — | — | 0.35 | |
| Net2DeeperNet (II)Training set size (S)=10002025.02 | 0.391 | — | — | 0.346 | |
| Baseline networkTraining set size (S)=10002025.02 | 0.391 | — | — | 0.351 | |
| LLT9bits=22026.03 | 0.396 | — | — | — | |
| Proposed (II)Training set size (S)=10002025.02 | 0.407 | — | — | 0.35 | |
| Prototype-based generative replay frameworkConfiguration=Best config, Learning mode=Incremental/Online2026.02 | 0.421 | — | 67.8 | — | |
| LLT4bits=22026.03 | 0.424 | — | — | — | |
| Forward ThinkingTraining set size (S)=10002025.02 | 0.43 | — | — | 0.38 | |
| Proposed (I)Training set size (S)=5002025.02 | 0.448 | — | — | 0.392 | |
| Random layer insertion (I)Training set size (S)=5002025.02 | 0.453 | — | — | 0.403 | |
| Proposed (II)Training set size (S)=5002025.02 | 0.455 | — | — | 0.398 | |
| Net2DeeperNet (II)Training set size (S)=5002025.02 | 0.455 | — | — | 0.407 | |
| Baseline networkTraining set size (S)=5002025.02 | 0.481 | — | — | 0.405 | |
| Forward ThinkingTraining set size (S)=5002025.02 | 0.541 | — | — | 0.44 | |
| Pr-MQbits=22026.03 | 0.593 | — | — | — | |
| Lasso-Pruned BaggingBase Estimator=Ridge, Active Est.=50, Comp. Ratio=0.0%, Latency / 1k (ms)=1.772026.06 | 0.6416 | — | 51.82 | — | |
| Standard BaggingBase Estimator=Ridge, Active Est.=50, Comp. Ratio=0.0%, Latency / 1k (ms)=2.112026.06 | 0.6451 | — | 51.57 | — | |
| SCSBBase Estimator=Ridge, Active Est.=10, Comp. Ratio=80.0%, Latency / 1k (ms)=0.51 (4.1×)2026.06 | 0.6504 | — | 51.16 | — | |
| ClimplicitProbe=Single-layer2025.04 | — | — | 44.2 | — | |
| CSPProbe=Single-layer2025.04 | — | — | 50.1 | — | |
| FS CHProbe=Single-layer2025.04 | — | — | 61.9 | — | |
| FS LocProbe=Single-layer2025.04 | — | — | 18.9 | — | |
| FS Loc + CHProbe=Single-layer2025.04 | — | — | 60.3 | — | |
| GeoCLIPProbe=Single-layer2025.04 | — | — | 58.9 | — | |
| Hierarchical MCTSBase Model=GPT-4.1-mini2025.11 | — | 21.59 | — | — | |
| Hierarchical MCTSBase LLM=GPT-4o2025.11 | — | 1,710 | — | — | |
| LATSBase Model=GPT-4.1-mini2025.11 | — | 0 | — | — | |
| LATSBase LLM=GPT-4o2025.11 | — | 65 | — | — | |
| MCTS-OutcomeBase Model=GPT-4.1-mini2025.11 | — | 0 | — | — | |
| MCTS-OutcomeBase LLM=GPT-4o2025.11 | — | 1,435 | — | — | |
| MCTS-ShapedBase Model=GPT-4.1-mini2025.11 | — | 21.59 | — | — | |
| MCTS-ShapedBase LLM=GPT-4o2025.11 | — | 1,159 | — | — | |
| NASTraining set size (S)=5002025.02 | — | — | — | 0.395 | |
| NASTraining set size (S)=10002025.02 | — | — | — | 0.347 | |
| ReActBase Model=GPT-4.1-mini2025.11 | — | 21.59 | — | — | |
| ReActBase LLM=GPT-4o2025.11 | — | 58 | — | — | |
| SatCLIPProbe=Single-layer2025.04 | — | — | 35.2 | — | |
| SINRProbe=Single-layer2025.04 | — | — | 37.1 | — | |
| TaxabindProbe=Single-layer2025.04 | — | — | 44.2 | — |