📘 200 AI Terms – Tanglish Version

A – Basics

1. AI (Artificial Intelligence) – Computer-ku manushan maari yosikka teach panradhu.
2. Machine Learning – Data-la irundhu self-ah kathukkara computer method.
3. Deep Learning – Neural network use panni periya periya pattern kathukkara ML.
4. Neural Network – Brain mathiri node-connect panni info process panradhu.
5. Algorithm – Oru problem solve panna step-by-step rules.
6. Model – Train pannina program prediction panna use aagum.
7. Training Data – Model-ku padikka kudukkara data.
8. Testing Data – Model padichadhu correct-ah iruka nu check panna data.
9. Validation Data – Training time tune panna use panra data.
10. Dataset – Collection of data arranged for AI padippu.

B – Learning Types

11. Supervised Learning – Input + answer kuduthu model train panradhu.
12. Unsupervised Learning – Answer illama patterns kandu pudikkara training.
13. Reinforcement Learning – Good action reward, wrong action penalty kuduthu train panradhu.
14. Classification – Data-va category-ku sort panradhu (spam vs non-spam).
15. Regression – Number predict panradhu (house price maari).
16. Clustering – Similar data-va group panradhu.
17. Anomaly Detection – Odd/abnormal data kandu pudikkaradhu.
18. Recommendation System – Suggest pannadhu (Netflix movies maari).
19. NLP (Natural Language Processing) – Computer-ku manush language puriyara madri padippu.
20. Computer Vision – Computer-ku image/video puriyara ability.

C – Words in Models

21. Token – Text-oda chinna unit (word or part of word).
22. Embedding – Text/image-a number-ah convert panna method.
23. Vector Database – Embedding store panna special DB.
24. RAG – Stored knowledge + new info combine pannitu answer kudukkara model.
25. Transformer – Innum pudhusaa irukkara AI architecture (GPT use pannathu).
26. Attention – Sentence-la important word focus panna technique.
27. LLM – Periya model, massive text-la train pannathu (ChatGPT maari).
28. Chatbot – Human-oda chat panna program.
29. Prompt – AI-ku kudukkara input text.
30. Prompt Engineering – Better prompt kuduthu nalla output vangara skill.

D – Training Related

31. Fine-Tuning – Already train pannina model-a namma data-la adjust panradhu.
32. Transfer Learning – Old model knowledge reuse panna new task-ku.
33. Pre-Trained Model – Already periya data-la padichirukkara model.
34. Zero-Shot – Example illama solve pannidradhu.
35. Few-Shot – Konjam example kuduthu solve pannadhu.
36. Generative AI – Pudhusaa text/image/music create pannidradhu.
37. Discriminative Model – Category separate panna model.
38. GAN – Oru model fake data create pannum, innoru detect pannum.
39. Diffusion Model – Step by step image generate pannidradhu.
40. Embedding Search – Closest match vector DB-la thedi pudikkaradhu.

E – Performance Terms

41. API – Software-ku AI-oda pesara interface.
42. Latency – AI answer kudukkara delay time.
43. Overfitting – Training data memorize panniduchu, new data-la fail aagum.
44. Underfitting – Model simple-ah irundhu data purinjuka mudiyala.
45. Bias – Training data wrong balance-a irundha unfair result.
46. Variance – Prediction too much vary aagardhu.
47. Accuracy – Correct prediction percentage.
48. Precision – Predict pannina positives-la correct irukardhu evlo.
49. Recall – Actual positives-la detect panninadhu evlo.
50. F1 Score – Precision + Recall balance measure.

F – Training Process

51. Epoch – Full training data one round pass panna time.
52. Batch Size – Oru time-la model process panna sample count.
53. Learning Rate – Model eppadi fast-ah kathukkudhu decide panna speed value.
54. Loss Function – Prediction wrong-a irundha error measure pannura formula.
55. Optimizer – Model error reduce panna adjustment method.
56. Gradient Descent – Step by step error kammi pannitu improve panra method.
57. Backpropagation – Neural network weight update panna reverse calculation.
58. Weights – Model decision influence panna internal number values.
59. Parameters – Model-ku learning-ku use aagura adjustable numbers.
60. Hyperparameters – Training start panna munadi set pannura settings.

G – Data Handling

61. Feature – Prediction-ku use panna input variable.
62. Label – Output or correct answer AI predict panna vendiyadhu.
63. Feature Engineering – Raw data-la irundhu useful inputs create panradhu.
64. Dimensionality Reduction – Data size kammi pannitu important info maintain panradhu.
65. PCA – Data-ku dimension kammi panna math method.
66. Normalization – Data value 0–1 range-ku scale pannuradhu.
67. Standardization – Data mean 0, variance 1 mathri normalize pannuradhu.
68. Data Augmentation – Training data variety create panna fake modifications (image flip maari).
69. Label Encoding – Text labels number-ku convert panradhu.
70. One-Hot Encoding – Category data binary 0/1 columns-la represent panradhu.

H – AI Transparency

71. Explainable AI – Human-ku AI decision puriya mathri explain panradhu.
72. Black Box Model – Inside logic puriyama irukkara model.
73. White Box Model – Step by step decision clear-a puriyum model.
74. AI Ethics – AI fair-a, safe-a build panna rules.
75. AI Safety – Human-ku harm panna koodatha AI build pannuradhu.
76. Hallucination – AI confident-ah wrong info create pannidradhu.
77. Token Limit – Model oru time process panna max text size.
78. Context Window – AI oru conversation-la remember pannidra text size.
79. Knowledge Graph – Facts + relation network representation.
80. Ontology – Concept + relationship formal representation.

I – Agents & Ops

81. AI Agent – Task autonomous-ah plan pannitu nadathidra program.
82. Multi-Agent System – Rendu moonu agents serndhu task solve panradhu.
83. Orchestration – Multiple models/agents manage panna process.
84. LLMOps – LLM manage panna production tools/process.
85. MLOps – ML models deployment + monitoring process.
86. Edge AI – Phone/IoT device-la AI run panradhu.
87. Cloud AI – Cloud servers-la AI run panradhu.
88. Federated Learning – Device-la data share pannama train panradhu.
89. Synthetic Data – Fake-a create panna but real-a maari irukkura training data.
90. Data Pipeline – Data collect → clean → train process path.

J – Data Prep

91. Data Cleaning – Wrong/miss data fix pannuradhu.
92. Data Labeling – Data-ku correct answer tag pannuradhu.
93. Ground Truth – Real correct output training-ku use pannuradhu.
94. Overparameterization – Too many parameters add pannudhu → overfit danger.
95. Pruning – Neural network-la extra parts remove pannitu fast panna.
96. Quantization – Model size compress panna precision reduce panradhu.
97. Knowledge Distillation – Small model big model-a copy pannitu train aagudhu.
98. Scaling Laws – Periya data + compute kudutha AI improve aagum-nu sollum rules.
99. AI Benchmark – Model performance test.
100. Open Source AI – Free/public ah use panna kudukkura AI tools.

K – Neural Network Varieties

101. CNN – Image data-ku special network.
102. RNN – Sequence/time-series data remember panna network.
103. LSTM – Long memory maintain panna RNN type.
104. GRU – Simple memory maintain panna RNN type.
105. Autoencoder – Data compress + reconstruct panna model.
106. VAE – Autoencoder type, new data generate panna use aagum.
107. Capsule Network – CNN advanced version, spatial relation maintain panradhu.
108. Encoder – Input purinjuka panna transformer part.
109. Decoder – Output generate panna transformer part.
110. Seq2Seq – Sequence input → sequence output model.

L – AI in Real Life

111. Speech Recognition – Voice convert panna text.
112. Text-to-Speech – Text convert panna human-like voice.
113. OCR – Image-la irukkura text read panna AI.
114. Chat Completion – Chat la human answer mathri output generate panna.
115. Image Generation – Prompt kuduthu pudhusaa image create panna.
116. Video Generation – Text description-la irundhu video create panna.
117. Code Generation – Program code AI auto write panradhu.
118. Style Transfer – Image content same, style mattum change panradhu.
119. Super Resolution – Image quality improve pannidradhu.
120. Speech Synthesis – Text → audio maari voice produce panradhu.

M – Probability Concepts

121. Bayesian Network – Probability-based decision model.
122. Markov Chain – Next state predict panna present state use pannidradhu.
123. Hidden Markov Model – Hidden states + sequence model.
124. Monte Carlo Simulation – Random sample use panni result estimate panna.
125. Probability Distribution – Outcome chances describe pannidradhu.
126. Prior Probability – Data munnaadi irukkura chance value.
127. Posterior Probability – Data apram update panna probability.
128. Likelihood – Model observed data match panna measure.
129. Entropy – Data uncertainty measure.
130. Information Gain – Data uncertainty reduce panna info.

N – Model Testing & Eval

131. Cross-Validation – Data split pannitu different parts-la model test panradhu.
132. Confusion Matrix – True vs predicted result table.
133. ROC Curve – Different thresholds-ku performance show pannura chart.
134. AUC – ROC curve area, performance score.
135. True Positive (TP) – Correct-ah positive predict pannadhu.
136. False Positive (FP) – Wrong-ah positive predict pannadhu.
137. True Negative (TN) – Correct-ah negative predict pannadhu.
138. False Negative (FN) – Miss pannina positive cases.
139. Cross-Entropy Loss – Classification-ku mostly use panna error measure.
140. MSE (Mean Squared Error) – Regression-ku error average.

O – Tools & Frameworks

141. TensorFlow – Google build pannina deep learning framework.
142. PyTorch – Meta build pannina famous deep learning tool.
143. Keras – Easy interface deep learning library.
144. Scikit-learn – Traditional ML-ku use panna Python library.
145. Hugging Face – Pretrained NLP models-ku famous platform.
146. LangChain – LLM apps build panna framework.
147. AutoML – AI models auto build panna toolset.
148. MLflow – ML experiments track + manage panna tool.
149. Weights & Biases – Training monitor panna SaaS tool.
150. ONNX – Model share panna open standard format.

P – Hardware & Compute

151. GPU – AI training fast panna graphic chip.
152. TPU – Google create pannina AI-special chip.
153. CPU – General computer processor.
154. FPGA – Customize panna hardware AI-ku use aagum.
155. ASIC – Special chip, oru task-ku design panninadhu.
156. Quantum AI – Quantum computer use panni AI run panradhu.
157. Distributed Training – Multiple machine-la parallel model train panradhu.
158. Parallel Processing – Same time-la multiple calculation nadakkaradhu.
159. Batch Processing – Large data chunk-by-chunk process panradhu.
160. Streaming Processing – Real-time continuous data process panradhu.

Q – Industry Uses

161. Autonomous Vehicle – AI-la run aagura self-driving car.
162. Healthcare AI – Disease detect panna medical AI usage.
163. FinTech AI – Banking/fraud detect panna AI.
164. E-commerce AI – Shopping site-la recommendation/chatbot AI.
165. Robotics – AI power kudutha machine action.
166. Smart Assistant – Siri/Alexa maari help pannura AI.
167. Predictive Maintenance – Machine breakdown munnaadi predict panna AI.
168. Sentiment Analysis – Text-la emotion understand panna AI.
169. EdTech AI – Student-ku personalized study kudukkura AI.
170. Gaming AI – Game NPC intelligent behaviour kudukkara AI.

R – Advanced Topics

171. Multimodal AI – Text + image + audio mix process panna AI.
172. Self-Supervised Learning – AI data itself use panni label generate panradhu.
173. Contrastive Learning – Similarity/difference compare pannitu model train panradhu.
174. Foundation Model – Periya general-purpose base model.
175. Autoregressive Model – Sequence next step predict panna model.
176. Masked Language Model – Sentence missing word predict panna AI.
177. Knowledge Transfer – Oru model padicha knowledge innoru task-ku transfer panradhu.
178. Curriculum Learning – Easy → hard order-la train panna method.
179. Chain of Thought (CoT) – Step by step reasoning AI-ku teach panradhu.
180. Tool-Using AI – AI external tools/APIs call panna capability.

S – Future & Governance

181. AGI – Human madri ella taskum pannidra intelligence.
182. ASI – Human-ai vida periya intelligence.
183. Narrow AI – Single task-ku specialist AI.
184. Ethical AI – Fair + bias illama irukkara AI.
185. Responsible AI – Accountable + safe AI build panradhu.
186. AI Governance – AI-ku policy + framework set pannadhu.
187. AI Regulation – Government rules AI usage-ku.
188. Fairness – Bias illama equal decision AI panna vendiyadhu.
189. Transparency – AI decision explain pannura clarity.
190. Trustworthy AI – Human nambikkai kidaikkara AI.

T – Buzzwords & Companies

191. AI Agent Framework – Agent build panna toolkit.
192. CrewAI – Multi-agent AI system framework.
193. Ollama – Local laptop-la LLM run panna tool.
194. RoboFlow – Computer vision dataset-ku tool.
195. Anthropic – Claude AI create pannina company.
196. OpenAI – ChatGPT create pannina company.
197. Google DeepMind – AlphaGo + research AI company.
198. Stability AI – Stable Diffusion create pannina company.
199. Perplexity AI – AI-powered search assistant.
200. Mistral AI – Open-source LLM build pannura company.