data/algorithms.json

{
  "algorithms": {
    "svm": {
      "name": "Support Vector Machine",
      "category": "classical_ml",
      "description": "A supervised learning algorithm that finds optimal hyperplanes for classification and regression tasks",
      "synonyms": [
        "support vector machine",
        "SVM",
        "support vector classifier",
        "support vector regression",
        "SVR"
      ],
      "blacklist": [
        "stroke volume monitoring",
        "severe viral meningitis",
        "syncope vasovagal mechanisms",
        "superior vena cava",
        "small vessel disease"
      ]
    },
    "decision_tree": {
      "name": "Decision Tree",
      "category": "classical_ml",
      "description": "A tree-like model that makes decisions by splitting data based on feature values",
      "synonyms": [
        "decision tree",
        "decision trees",
        "DT",
        "CART",
        "classification tree",
        "regression tree"
      ]
    },
    "random_forest": {
      "name": "Random Forest",
      "category": "classical_ml",
      "description": "An ensemble method that combines multiple decision trees for improved accuracy",
      "synonyms": [
        "random forest",
        "RF",
        "random forests",
        "forest classifier"
      ],
      "blacklist": [
        "radiofrequency",
        "rheumatoid factor",
        "risk factor",
        "renal failure",
        "respiratory failure",
        "reticular formation"
      ]
    },
    "xgboost": {
      "name": "XGBoost",
      "category": "classical_ml",
      "description": "Extreme Gradient Boosting - an optimized gradient boosting framework",
      "synonyms": [
        "xgboost",
        "XGBoost",
        "extreme gradient boosting",
        "XGB"
      ]
    },
    "logistic_regression": {
      "name": "Logistic Regression",
      "category": "classical_ml",
      "description": "A linear model for binary and multiclass classification problems",
      "synonyms": [
        "logistic regression",
        "logit",
        "logistic model",
        "LR"
      ]
    },
    "naive_bayes": {
      "name": "Naive Bayes",
      "category": "classical_ml",
      "description": "A probabilistic classifier based on Bayes' theorem with independence assumptions",
      "synonyms": [
        "naive bayes",
        "Naive Bayes",
        "NB",
        "Bayes classifier"
      ]
    },
    "knn": {
      "name": "K-Nearest Neighbors",
      "category": "classical_ml",
      "description": "A non-parametric method that classifies data points based on the class of their nearest neighbors",
      "synonyms": [
        "k-nearest neighbors",
        "KNN",
        "k-NN",
        "nearest neighbor",
        "k nearest neighbour"
      ]
    },
    "kmeans": {
      "name": "K-Means Clustering",
      "category": "classical_ml",
      "description": "An unsupervised clustering algorithm that partitions data into k clusters",
      "synonyms": [
        "k-means",
        "K-means",
        "kmeans",
        "k-means clustering",
        "k means"
      ]
    },
    "gradient_boosting": {
      "name": "Gradient Boosting",
      "category": "classical_ml",
      "description": "An ensemble method that builds models sequentially to correct errors of previous models",
      "synonyms": [
        "gradient boosting",
        "GB",
        "GBM",
        "gradient boosted trees",
        "gradient boosting machine"
      ]
    },
    "ada_boost": {
      "name": "AdaBoost",
      "category": "classical_ml",
      "description": "Adaptive Boosting algorithm that combines weak learners into a strong classifier",
      "synonyms": [
        "AdaBoost",
        "ada boost",
        "adaptive boosting",
        "adaboost"
      ]
    },
    "pca": {
      "name": "Principal Component Analysis",
      "category": "classical_ml",
      "description": "A dimensionality reduction technique that finds principal components of data variance",
      "synonyms": [
        "PCA",
        "principal component analysis",
        "principal components"
      ],
      "blacklist": [
        "posterior cerebral artery",
        "prostate cancer antigen",
        "patient-controlled analgesia",
        "percutaneous coronary angioplasty",
        "primary care physician",
        "polycystic ovary syndrome"
      ]
    },
    "linear_regression": {
      "name": "Linear Regression",
      "category": "classical_ml",
      "description": "A linear approach to modeling the relationship between variables",
      "synonyms": [
        "linear regression",
        "ordinary least squares",
        "OLS",
        "linear model"
      ]
    },
    "cnn": {
      "name": "Convolutional Neural Network",
      "category": "deep_learning",
      "description": "Deep learning architecture specialized for processing grid-like data such as images",
      "synonyms": [
        "convolutional neural network",
        "CNN",
        "ConvNet",
        "convolutional network",
        "deep convolutional",
        "conv neural network",
        "convolution neural network"
      ],
      "blacklist": [
        "cranial nerve nuclei",
        "central nervous system",
        "chronic kidney disease",
        "clinical nurse navigator",
        "calcineurin inhibitor"
      ]
    },
    "lstm": {
      "name": "Long Short-Term Memory",
      "category": "deep_learning",
      "description": "A type of recurrent neural network capable of learning long-term dependencies",
      "synonyms": [
        "LSTM",
        "long short-term memory",
        "long short term memory",
        "LSTM network"
      ]
    },
    "transformer": {
      "name": "Transformer",
      "category": "deep_learning",
      "description": "Attention-based neural network architecture for sequence-to-sequence tasks",
      "synonyms": [
        "transformer",
        "transformers",
        "transformer model",
        "transformer architecture",
        "self-attention"
      ]
    },
    "resnet": {
      "name": "ResNet",
      "category": "deep_learning",
      "description": "Residual Neural Network - a deep CNN architecture with skip connections",
      "synonyms": [
        "ResNet",
        "resnet",
        "residual network",
        "residual neural network"
      ]
    },
    "unet": {
      "name": "U-Net",
      "category": "deep_learning",
      "description": "A CNN architecture for biomedical image segmentation with encoder-decoder structure",
      "synonyms": [
        "U-Net",
        "UNet",
        "u-net",
        "unet"
      ]
    },
    "gan": {
      "name": "Generative Adversarial Network",
      "category": "deep_learning",
      "description": "A framework where two neural networks compete to generate realistic data",
      "synonyms": [
        "generative adversarial network",
        "generative adversarial networks",
        "GANs",
        "GAN model",
        "GAN network",
        "adversarial network",
        "adversarial training"
      ],
      "blacklist": [
        "giant axonal neuropathy",
        "Giant Axonal Neuropathy",
        "GAN neuropathy",
        "axonal neuropathy",
        "ganglion",
        "ganglia",
        "ganglioside",
        "gangliosides",
        "ganglionic",
        "gangrene",
        "gangrenous",
        "Ganoderma",
        "ganoderic",
        "ganciclovir",
        "gastric antral nodularity",
        "gonadotropin-releasing hormone antagonist",
        "glucosamine",
        "galactosamine",
        "N-acetylgalactosamine",
        "GalNAc"
      ]
    },
    "autoencoder": {
      "name": "Autoencoder",
      "category": "deep_learning",
      "description": "Neural networks that learn efficient representations by encoding and decoding data",
      "synonyms": [
        "autoencoder",
        "auto-encoder",
        "autoencoders",
        "variational autoencoder",
        "VAE"
      ]
    },
    "vgg": {
      "name": "VGG",
      "category": "deep_learning",
      "description": "Very Deep Convolutional Networks - a CNN architecture with small convolution filters",
      "synonyms": [
        "VGG",
        "VGG-16",
        "VGG-19",
        "VGGNet"
      ]
    },
    "rnn": {
      "name": "Recurrent Neural Network",
      "category": "deep_learning",
      "description": "Neural networks with memory that can process sequences of data",
      "synonyms": [
        "RNN",
        "recurrent neural network",
        "recurrent network",
        "RNNs"
      ],
      "blacklist": [
        "ribonuclease",
        "registered nurse navigator",
        "reactive nitrogen species"
      ]
    },
    "gru": {
      "name": "Gated Recurrent Unit",
      "category": "deep_learning",
      "description": "Simplified variant of LSTM with fewer parameters and faster training",
      "synonyms": [
        "GRU",
        "gated recurrent unit",
        "gated recurrent units",
        "GRUs"
      ]
    },
    "yolo": {
      "name": "YOLO",
      "category": "deep_learning",
      "description": "You Only Look Once - real-time object detection algorithm",
      "synonyms": [
        "YOLO",
        "you only look once",
        "YOLOv3",
        "YOLOv4",
        "YOLOv5"
      ]
    },
    "capsnet": {
      "name": "Capsule Network",
      "category": "deep_learning",
      "description": "Neural network architecture that uses capsules to better model hierarchical relationships",
      "synonyms": [
        "CapsNet",
        "capsule network",
        "capsule networks",
        "dynamic routing"
      ]
    },
    "gpt": {
      "name": "GPT",
      "category": "llms",
      "description": "Generative Pre-trained Transformer - OpenAI's large language model family",
      "synonyms": [
        "GPT",
        "gpt",
        "generative pre-trained transformer",
        "ChatGPT",
        "GPT-3",
        "GPT-4",
        "GPT-4o",
        "OpenAI GPT"
      ],
      "blacklist": [
        "glucose-6-phosphate transporter",
        "glutamic pyruvic transaminase",
        "glutathione peroxidase",
        "glycerophosphate",
        "guanosine triphosphate"
      ]
    },
    "claude": {
      "name": "Claude",
      "category": "llms",
      "description": "Anthropic's AI assistant and large language model family",
      "synonyms": [
        "Claude",
        "claude",
        "Anthropic Claude",
        "Claude-3",
        "Claude Sonnet",
        "Claude Haiku",
        "Claude Opus"
      ]
    },
    "bert": {
      "name": "BERT",
      "category": "llms",
      "description": "Bidirectional Encoder Representations from Transformers - Google's pre-trained language model",
      "synonyms": [
        "BERT",
        "bert",
        "bidirectional encoder representations",
        "BERT model",
        "Google BERT"
      ],
      "blacklist": [
        "behavioral emergency response team",
        "biomedical emergency response team",
        "blood-retinal barrier transport",
        "bronchial epithelial cell"
      ]
    },
    "gemini": {
      "name": "Gemini",
      "category": "llms",
      "description": "Google's multimodal large language model family",
      "synonyms": [
        "Gemini",
        "gemini",
        "Google Gemini",
        "Gemini Pro",
        "Gemini Ultra",
        "Gemini Nano"
      ]
    },
    "llama": {
      "name": "LLaMA",
      "category": "llms",
      "description": "Large Language Model Meta AI - Meta's open-source language model family",
      "synonyms": [
        "LLaMA",
        "llama",
        "Llama",
        "Meta LLaMA",
        "Llama-2",
        "Llama 2",
        "Llama 3",
        "Code Llama"
      ]
    },
    "qwen": {
      "name": "Qwen",
      "category": "llms",
      "description": "Alibaba's large language model series with multilingual capabilities",
      "synonyms": [
        "Qwen",
        "qwen",
        "Alibaba Qwen",
        "Qwen-7B",
        "Qwen-14B",
        "Qwen-72B",
        "Tongyi Qianwen"
      ]
    },
    "deepseek": {
      "name": "DeepSeek",
      "category": "llms",
      "description": "DeepSeek's large language model optimized for code and reasoning",
      "synonyms": [
        "DeepSeek",
        "deepseek",
        "DeepSeek Coder",
        "DeepSeek LLM",
        "DeepSeek-V2"
      ]
    },
    "mistral": {
      "name": "Mistral",
      "category": "llms",
      "description": "Mistral AI's efficient and powerful open-source language models",
      "synonyms": [
        "Mistral",
        "mistral",
        "Mistral 7B",
        "Mixtral",
        "Mistral AI",
        "Mixtral 8x7B"
      ]
    },
    "palm": {
      "name": "PaLM",
      "category": "llms",
      "description": "Pathways Language Model - Google's large-scale language model",
      "synonyms": [
        "PaLM",
        "palm",
        "Pathways Language Model",
        "PaLM-2",
        "Google PaLM"
      ]
    },
    "t5": {
      "name": "T5",
      "category": "llms",
      "description": "Text-to-Text Transfer Transformer - Google's unified text processing model",
      "synonyms": [
        "T5",
        "t5",
        "text-to-text transfer transformer",
        "Google T5"
      ],
      "blacklist": [
        "T5 vertebra",
        "T5 spinal",
        "fifth thoracic vertebra",
        "thoracic vertebra 5",
        "T5 nerve root",
        "T5 dermatome"
      ]
    },
    "roberta": {
      "name": "RoBERTa",
      "category": "llms",
      "description": "Robustly Optimized BERT Pretraining Approach - Meta's improved BERT variant",
      "synonyms": [
        "RoBERTa",
        "roberta",
        "robustly optimized BERT",
        "Meta RoBERTa"
      ]
    },
    "phi": {
      "name": "Phi",
      "category": "llms",
      "description": "Microsoft's small language model series optimized for efficiency",
      "synonyms": [
        "Microsoft Phi",
        "Microsoft Phi-3",
        "Microsoft Phi-2", 
        "Microsoft Phi-1",
        "Phi language model",
        "Phi LLM",
        "Phi small language model"
      ],
      "blacklist": [
        "phi angle",
        "phi coefficient",
        "phi correlation",
        "dihedral angle",
        "phi psi angles",
        "ramachandran plot",
        "protein phi",
        "phi torsion",
        "golden ratio phi",
        "phi statistic",
        "phi phenomenon",
        "phi value analysis",
        "protected health information",
        "personal health information",
        "phosphatidylinositol",
        "bacteriophage phi",
        "phage phi",
        "phi bacteriophage",
        "phi X174",
        "phi 6",
        "phi 29",
        "phi92",
        "magnetic flux phi",
        "flux phi",
        "phi magnetic",
        "volumetric flux",
        "flow rate phi",
        "phi value",
        "phi values",
        "phi analysis",
        "phi-value analysis",
        "protein folding phi",
        "transition state phi",
        "nucleation condensation phi",
        "pharmacologic MRI phi",
        "phMRI",
        "DICOM phi",
        "medical imaging phi",
        "phi function",
        "phi distribution",
        "phi parameter",
        "phi variable",
        "phi measurement",
        "phi calculation"
      ]
    },
    "falcon": {
      "name": "Falcon",
      "category": "llms",
      "description": "Technology Innovation Institute's open-source large language model",
      "synonyms": [
        "Falcon",
        "falcon",
        "Falcon-7B",
        "Falcon-40B",
        "Falcon-180B",
        "TII Falcon"
      ]
    }
  }
}