A comprehensive bilingual glossary covering 50 fundamental artificial intelligence terms — from core concepts like machine learning and neural networks to specialized vocabulary in generative AI, cybersecurity, fintech, defense, and business strategy. Each term explained in Turkish with English context for cross-border AI practitioners.
Section 1: Core Concepts (5 Terms)
Artificial Intelligence (Yapay Zeka): The broad field of computer science focused on creating systems capable of performing tasks that typically require human intelligence — reasoning, learning, perception, and decision-making. In Turkey's context, AI encompasses the 457 startups and $1B market driving the country's technology transformation. Machine Learning (Makine Öğrenmesi): A subset of AI where algorithms improve through experience rather than explicit programming. Machine learning models identify patterns in data to make predictions or decisions. This is the foundation of most commercial AI applications in Turkey's startup ecosystem. Deep Learning (Derin Öğrenme): A specialized form of machine learning using neural networks with multiple layers to process complex data representations. Deep learning powers image recognition, speech processing, and natural language understanding systems. Neural Network (Sinir Ağı): Computing architectures inspired by biological neural networks, consisting of interconnected nodes organized in layers. Neural networks form the computational backbone of modern AI systems, from simple classification to complex generative models. Natural Language Processing (Doğal Dil İşleme): The AI discipline focused on enabling machines to understand, interpret, and generate human language. NLP is particularly significant for Turkish AI, given the agglutinative nature of the Turkish language that presents unique computational challenges.
Section 2: Models and Training (5 Terms)
Large Language Model / LLM (Büyük Dil Modeli): AI models trained on massive text corpora that can generate, translate, summarize, and analyze text with human-level fluency. GPT, Claude, and Gemini are prominent examples. Turkish-language LLMs represent a growing research frontier. Transfer Learning (Transfer Öğrenme): A technique where a model trained on one task is adapted for a different but related task. Transfer learning dramatically reduces the data and compute required to build specialized AI applications, making it crucial for resource-constrained Turkish startups. Fine-Tuning (İnce Ayar): The process of taking a pre-trained model and further training it on domain-specific data. Fine-tuning enables general-purpose models to achieve specialist-level performance in fields like Turkish legal text analysis or medical diagnostics. Training Data (Eğitim Verisi): The dataset used to teach a machine learning model to recognize patterns and make predictions. Training data quality directly determines model performance — a critical consideration in Turkish AI where high-quality Turkish-language datasets remain scarce. Overfitting (Aşırı Uyum): A modeling error where an algorithm learns the training data too precisely, including its noise and outliers, resulting in poor performance on new, unseen data. Overfitting is a persistent challenge in AI development, requiring regularization techniques and validation protocols.
Section 3: Generative AI (5 Terms)
Generative AI (Üretken Yapay Zeka): AI systems that create new content — text, images, audio, video, or code — rather than simply analyzing existing data. Generative AI represents the fastest-growing segment of the global AI market and underpins several OSP portfolio company technologies. Prompt Engineering (Komut Mühendisliği): The practice of crafting precise instructions to guide generative AI models toward desired outputs. Prompt engineering has emerged as a distinct professional skill, combining linguistic precision with understanding of model behavior and limitations. Hallucination (Halüsinasyon): When an AI model generates factually incorrect or fabricated information presented as truth. Hallucination remains one of the most significant challenges in deploying generative AI for business-critical applications, particularly in regulated domains like healthcare and finance. Retrieval-Augmented Generation / RAG (Erişim Destekli Üretim): An architecture that combines generative AI with information retrieval systems, grounding model outputs in verified source documents. RAG significantly reduces hallucination and is essential for enterprise AI deployments requiring factual accuracy. Embedding (Gömme Vektörü): Mathematical representations that convert text, images, or other data into numerical vectors capturing semantic meaning. Embeddings enable AI systems to measure similarity, perform search, and cluster related content — foundational infrastructure for RAG systems and recommendation engines.
Section 4: Automation and Agents (5 Terms)
AI Agent (Yapay Zeka Ajanı): An autonomous AI system capable of perceiving its environment, making decisions, and executing multi-step actions to achieve defined objectives. AI agents represent the next evolution beyond chatbots — systems that can independently complete complex workflows. Autonomous System (Otonom Sistem): Technology that operates independently without continuous human control. In the Turkish context, autonomous systems span from self-driving vehicles and delivery drones to industrial robots and defense platforms. Turkey's indigenous autonomous drone program has positioned the country as a global leader. Robotic Process Automation / RPA (Robotik Süreç Otomasyonu): Software that automates repetitive, rule-based digital tasks — data entry, form processing, report generation. RPA represents the most accessible entry point for organizations beginning their AI transformation journey, with rapid ROI and minimal technical complexity. MLOps (Makine Öğrenmesi Operasyonları): The practices and tools for deploying, monitoring, and maintaining machine learning models in production environments. MLOps bridges the gap between model development and real-world deployment — a discipline where many Turkish AI startups need strengthening. API — Application Programming Interface (Uygulama Programlama Arayüzü): Standardized interfaces enabling different software systems to communicate. APIs are the connective tissue of the AI ecosystem, allowing startups to integrate third-party AI capabilities and expose their own models as services.
Section 5: Data and Analytics (5 Terms)
Big Data (Büyük Veri): Datasets too large or complex for traditional processing methods, characterized by volume, velocity, variety, veracity, and value. Turkey's 85-million population and digitizing economy generate massive data flows that fuel AI model training and business intelligence. Data Mining (Veri Madenciliği): The process of discovering patterns, correlations, and anomalies in large datasets using statistical and machine learning methods. Data mining transforms raw information into actionable business intelligence — a capability increasingly demanded by Turkish enterprises undergoing digital transformation. Feature Engineering (Özellik Mühendisliği): The process of selecting, transforming, and creating input variables that improve machine learning model performance. Feature engineering remains one of the most impactful yet underappreciated skills in applied AI, often determining the difference between mediocre and exceptional model accuracy. Data Labeling (Veri Etiketleme): The process of annotating raw data with informative tags that enable supervised machine learning. Data labeling is labor-intensive and represents a significant cost center in AI development. Turkey's competitive labor costs make it an attractive location for data labeling operations serving global AI companies. Data Lake (Veri Gölü): A centralized repository storing structured and unstructured data at scale in its raw format. Data lakes provide the flexible storage infrastructure that AI systems require, enabling organizations to retain all data and apply analytical frameworks as needs evolve.
Section 6: Computer Vision (5 Terms)
Computer Vision (Bilgisayarlı Görü): The AI field enabling machines to interpret and understand visual information from images and video. Computer vision applications range from industrial quality inspection and autonomous navigation to medical imaging diagnostics and agricultural monitoring. Object Detection (Nesne Algılama): The computer vision task of identifying and localizing specific objects within images or video frames. Object detection powers applications from autonomous driving (detecting pedestrians and vehicles) to retail analytics (tracking customer behavior) and security surveillance. OCR — Optical Character Recognition (Optik Karakter Tanıma): Technology that converts images of text into machine-readable digital text. OCR is particularly relevant for Turkish and Arabic language processing, where complex scripts and diacritical marks present recognition challenges that AI-enhanced OCR systems increasingly overcome. GAN — Generative Adversarial Network (Üretici Çekişmeli Ağ): A deep learning architecture where two neural networks compete — one generating synthetic data, the other evaluating authenticity. GANs produce remarkably realistic images, video, and audio, with applications in creative industries, data augmentation, and simulation. Image Segmentation (Görüntü Segmentasyonu): The process of partitioning an image into meaningful regions at the pixel level. Segmentation enables precise understanding of visual scenes — distinguishing tumor boundaries in medical scans, identifying crop zones in satellite imagery, or isolating objects for augmented reality applications.
Section 7: Cybersecurity AI (5 Terms)
Anomaly Detection (Anomali Tespiti): AI techniques that identify patterns deviating from established baselines in network traffic, user behavior, or system operations. Anomaly detection forms the first line of AI-powered cyber defense, catching novel threats that signature-based systems miss. Threat Intelligence (Tehdit İstihbaratı): The collection, analysis, and dissemination of information about current and emerging cyber threats. AI-powered threat intelligence platforms process millions of indicators of compromise in real time, providing security teams with actionable insights about attack vectors targeting their specific industry and geography. SIEM — Security Information and Event Management (Güvenlik Bilgi ve Olay Yönetimi): Platforms that aggregate, correlate, and analyze security event data from across an organization's technology infrastructure. AI-enhanced SIEM systems reduce alert fatigue by filtering false positives and prioritizing genuine threats based on contextual risk assessment. Zero Trust (Sıfır Güven): A security architecture that eliminates implicit trust and requires continuous verification of every user, device, and network flow. Zero trust has become the dominant cybersecurity paradigm, particularly relevant for Gulf organizations undergoing rapid digital transformation. Penetration Testing (Penetrasyon Testi): Authorized simulated cyberattacks conducted to evaluate system security. AI is transforming penetration testing from manual, periodic assessments to continuous automated vulnerability discovery — a service offered by several Turkish cybersecurity firms including OSP portfolio companies.
Section 8: Fintech AI (5 Terms)
Algorithmic Trading (Algoritmik Ticaret): Automated trading systems that execute transactions based on predefined rules and AI-driven market analysis. Algorithmic trading now accounts for the majority of global equity trading volume, with Turkish fintech startups developing systems optimized for Borsa Istanbul and regional exchanges. Credit Scoring (Kredi Skorlama): AI models that assess borrower creditworthiness by analyzing traditional financial data alongside alternative signals — social media activity, transaction patterns, device data. AI-powered credit scoring expands financial inclusion by evaluating individuals and businesses that traditional scoring systems cannot assess. Robo-Advisor (Robo-Danışman): Automated investment management platforms that use algorithms to construct, rebalance, and optimize portfolios based on individual risk profiles. Robo-advisors democratize wealth management, reducing minimum investment thresholds from thousands to single dollars. RegTech — Regulatory Technology (Düzenleyici Teknoloji): AI-powered solutions that automate compliance monitoring, regulatory reporting, and risk management. RegTech is critical in highly regulated Gulf financial markets, where compliance requirements are expanding faster than manual processes can accommodate. AML — Anti-Money Laundering (Kara Para Aklama ile Mücadele): AI systems that detect suspicious financial transactions indicative of money laundering, terrorist financing, or sanctions evasion. AML compliance is a universal requirement for financial institutions, and AI dramatically improves detection accuracy while reducing false positive rates that burden compliance teams.
Section 9: Defense AI (5 Terms)
Swarm Intelligence (Sürü Zekası): Coordination algorithms enabling multiple autonomous agents to operate collectively, mimicking the emergent behavior of biological swarms. Turkey has demonstrated swarm drone capabilities that position the country among the global leaders in this militarily significant technology. Autonomous Drone (Otonom Drone): Unmanned aerial vehicles capable of executing missions without continuous human control. Turkey's Bayraktar and Anka platforms have established the country's reputation in autonomous aviation, spawning an ecosystem of AI startups building supporting technologies — navigation, target recognition, and mission planning. C4ISR — Command, Control, Communications, Computers, Intelligence, Surveillance, and Reconnaissance: The integrated military information system architecture that AI is fundamentally transforming. AI-enabled C4ISR provides real-time battlefield awareness, predictive threat assessment, and automated decision support that compresses the observe-orient-decide-act loop. Electronic Warfare (Elektronik Harp): The use of electromagnetic spectrum for military advantage — jamming, spoofing, and intercepting enemy communications and radar. AI enhances electronic warfare through adaptive signal processing, cognitive radar, and autonomous spectrum management. Tactical AI (Taktik Yapay Zeka): AI systems designed for real-time decision support in operational military environments. Tactical AI processes sensor data, terrain information, and threat assessments to recommend courses of action — augmenting human commanders with computational speed and pattern recognition capabilities.
Section 10: Business and Strategy (5 Terms)
AI Maturity (Yapay Zeka Olgunluğu): A framework assessing an organization's readiness and capability to effectively deploy and benefit from AI technologies. AI maturity models evaluate data infrastructure, talent, processes, culture, and governance — providing a roadmap for progressive AI adoption that OSP uses in consulting engagements. Digital Transformation (Dijital Dönüşüm): The comprehensive integration of digital technology into all areas of business operations, fundamentally changing how organizations deliver value. Digital transformation provides the foundational infrastructure — data systems, cloud computing, APIs — upon which AI capabilities are built. ROI — Return on Investment (Yatırım Getirisi): The financial metric measuring the profitability of an investment relative to its cost. Quantifying AI ROI remains challenging due to indirect benefits — productivity gains, error reduction, faster decision-making — that traditional financial models struggle to capture. CAGR — Compound Annual Growth Rate (Bileşik Yıllık Büyüme Oranı): The annualized rate of return that smooths growth over multiple periods. CAGR is essential for evaluating AI market projections — Turkey's AI market is growing at approximately 35 percent CAGR, outpacing most European markets. TAM/SAM/SOM — Total Addressable Market, Serviceable Addressable Market, Serviceable Obtainable Market (Toplam Adreslenebilir Pazar / Hizmet Verilebilir Pazar / Elde Edilebilir Pazar): The three-tier market sizing framework used in startup valuation. Understanding TAM/SAM/SOM is critical for investors evaluating Turkish AI startups — distinguishing between the global AI opportunity, the Turkish-addressable portion, and the realistically capturable market share.