Unverified50% confidenceBenchmarkExact time
Under the single-step inference setting, BCSM reduces the number of neural function evaluations (NFE) by up to 6 times compared with baselines, while achieving a 16.9% relative improvement on HumanEval.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/4/2026
First Seen
Valid until: 2/2/2027
Sources
Related Entities
Related Claims
UnverifiedCompared with the most efficient baseline, the metacognitive controller reduces inference FLOPs by an average of 20.6% while maintaining competitive task accuracy.67% similarUnverifiedA compact spiking neural network with 93,000 parameters has an inference energy consumption 15 times lower than the reference solution, at the cost of a 0.073 decrease in macro F1.67% similarUnverifiedIn the application to neural-network correction of emission inventories, the error of biogenic source correction was reduced by a factor of 2.4–5, and low-temperature emission information that a single inversion would lose was recovered.66% similarUnverifiedIn half of the tested scenarios, the thresholds selected by neural models and BKT differed by more than 0.10.65% similarVerifiedMany neural networks exhibit an overconfidence calibration bias, where output confidence scores remain close to 1 even when the error rate is high.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1133924API
curl https://kongchang.com/api/v1/knowledge/claims/1133924MCP
get_claim(id=1133924)