Kernel Methods TRIZ Cases
- Neural Network Accelerator Resource Utilization via Feature Map Slicing
Configures hardware accelerators for convolutional neural networks using fine-grained row and column buffering with pipelined data slices.
- Consensus-Based Neural Network Sizing for IoT Efficiency
A consensus mechanism determines optimal neural network complexity through progressive inference agreement.
- Causal Invariance Feature Selection for Robust Prediction
A prediction model selects features with causal invariance to ensure accurate results across varying environments.
- Event Tweet Collection via Keyword Clustering and Active Learning
System segments tweet collection into stages, filtering noise through clustering to resolve precision-volume trade-offs.
- Automated Video Ratings Using Feature Sets
Automated video ratings use feature sets to assign maturity labels, replacing manual review with digital fingerprinting for throughput.
- Anomaly Detection via Low-Dimensional Feature Subsets
Segments high-dimensional data into low-dimensional subspaces to detect and classify anomalous clusters, reducing manual examination time.
- Automated Data Structuring via Machine Learning for Software Testing
Machine learning engine analyzes log files to generate structured datasets, reducing manual testing time and improving software validation throughput.
- Autoencoder Clustering for Network Anomaly Detection
Autoencoder clustering segments variable network traffic to reduce false positives while maintaining high anomaly detection sensitivity.
- ESOL Script Assessment via Discriminative Ranking
Discriminative models assess linguistic competence while resisting prompt-specific subversion attempts.
- Zero-Shot Learning Device Using Correspondence Inference
A learning device infers correspondences between seen task outputs and unseen class attributes to generate predictions without new model construction.
- Analytic Team Bias Detection in Immersive Environment
Problem analysis system detects bias conditions via sensor data and adjusts immersive environment variables to improve team decision accuracy.
- Optimum Sampling Search System with Risk Assessment
An optimum sampling search system calculates constraint satisfaction probabilities to recommend and adjust sampling parameters.
- Automated Research System for Biomedical Experimentation
An automated research system accelerates biomedical experimentation through modular robotic platforms and integrated data processing engines.
- Adaptive Thermal Diffusivity Kernel for Additive Manufacturing Phase Detection
Predicting adaptive thermal diffusivity kernel via machine learning resolves voxel temperature ambiguity, improving geometrical accuracy.
- Homomorphic Encryption Compatible Machine Learning Model Training
A machine learning model modified with polynomial functions compatible with homomorphic encryption for secure third-party training.
- SVM Classification Using Segmented RBF Kernels
A classification system segments data sets and selects radial basis function kernels for support vector machines.
- Abnormal Behavior Detection Model Using Word Vector Clustering
A detection model building apparatus clusters word vectors from program operation sequences to classify malicious behavior patterns.
- Automated News Ranking for Financial Analysis
Entity-based tracking and clustering models filter noise to surface critical financial news without manual subscription management.
- MRI Coil Fault Detection Using Pre-Scan Reference Signals
Automated fault detection using pre-scan reference signals reduces diagnostic time and user workload without requiring image acquisition.
- Cognitive Test Familiarity Estimation Device
A determination device estimates required trial quantities based on subject attributes to assess test familiarity.
- Deep Tensor Column Execution for Neural Network Cache Optimization
Partitioning neural network execution into sub-tensor columns reduces memory access and redundant calculations by keeping intermediate values in cache.
- In-Memory Convolution Accelerator for CNN Data Movement Reduction
An in-memory computation array stores kernel matrix elements to perform parallel convolution operations directly within memory cells.
- Biological System State Prediction Using Neural Network Feature Extraction
Encoder-decoder architecture processes microscope images and metadata to predict biological system health, enabling non-invasive risk parameter identification.
- Response Phrase Selection for Ogiri Conversational Art
A response phrase selection device calculates scores based on theme phrase feature values to identify appropriate conversational replies.