Training of DHN Part 1

Duration: 12 min

This video lesson is available to enrolled students.

Return to /learn/NTA-UGC-NET-PAPER-2/unit-10-artificial-intelligence/artificial-intelligence-artificial-neural-networks/training-of-hopfield-networks/asset-training-of-dhn-part-1 after enrolling

Inside: a video lesson and guided study material.

Module outline

  1. Course Overview: About the Course
  2. Paper - 1 | Unit - 1 | Teaching Aptitude: Concept and Objectives of Teaching / शिक्षण की अवधारणा एवं उद्देश्य, Levels of Teaching / शिक्षण के स्तर, Learner's Characteristics / शिक्षार्थी की विशेषताएँ, Learning: Nature and Theories / अधिगम: प्रकृति एवं सिद्धांत, Cognition, Memory and Metacognition / संज्ञान, स्मृति एवं अधिसंज्ञान, Factors Affecting Teaching / शिक्षण को प्रभावित करने वाले कारक, Effective Teaching and the Teacher / प्रभावी शिक्षण एवं शिक्षक, Instructional Objectives and Lesson Planning / अनुदेशनात्मक उद्देश्य एवं पाठ नियोजन, Methods of Teaching / शिक्षण विधियाँ, Offline and Online Teaching Methods / ऑफ़लाइन एवं ऑनलाइन शिक्षण विधियाँ, Teaching Support System / शिक्षण सहायक प्रणाली, Evaluation Systems / मूल्यांकन प्रणाली, Biological Basis of Emotion / संवेग का जैविक आधार, Central and Peripheral Nervous Systems / केन्द्रीय और परिधीय तंत्रिका तंत्र
  3. Paper - 1 | Unit - 2 | Research Aptitude: Research: Meaning, Features and Scientific Method / अनुसंधान: अर्थ, लक्षण व वैज्ञानिक विधि, Research Paradigms and Approaches / अनुसंधान के प्रतिमान एवं उपागम, Types of Research / अनुसंधान के प्रकार, Variables and Measurement / चर एवं मापन, Methods of Research / अनुसंधान की विधियाँ, Steps of Research / अनुसंधान के चरण, Sampling / प्रतिचयन, Tools and Methods of Data Collection / आँकड़ा संग्रह के उपकरण एवं विधियाँ, Data Analysis and Statistics in Research / आँकड़ा विश्लेषण एवं शोध सांख्यिकी, Thesis and Article Writing / शोध-प्रबंध एवं लेख लेखन, Application of ICT in Research / अनुसंधान में ICT का अनुप्रयोग, Research Ethics / शोध नैतिकता
  4. Paper - 1 | Unit - 3 | Comprehension: Literal Comprehension, Vocabulary and Language in Context, Inferential Comprehension, Critical Comprehension, Comprehension / Reading Comprehension / Unseen Passages (Critical Reasoning) (Paragraph Questions)
  5. Paper - 1 | Unit - 4 | Communication: Communication: Meaning, Nature and Process / संचार: अर्थ, प्रकृति एवं प्रक्रिया, Types and Levels of Communication / संचार के प्रकार एवं स्तर, Verbal and Non-verbal Communication / मौखिक एवं अमौखिक संचार, Semiotics and Meaning / संकेत विज्ञान एवं अर्थ, Models and Theories of Communication / संचार के मॉडल एवं सिद्धांत, Effective Communication / प्रभावी संचार, Barriers to Communication / संचार की बाधाएँ, Classroom Communication / कक्षा संचार, Mass Media: Types and Evolution / जनसंचार माध्यम: प्रकार एवं विकास, Mass Media and Society / जनसंचार एवं समाज, Digital and New Media / डिजिटल एवं नव माध्यम, Morphology and Syntax, Public Health Nutrition / जन स्वास्थ्य पोषण, Newtonian Motion and Stability, GK- Current Affairs / GK- समसामयिकी, Indian Economy / भारतीय अर्थव्यवस्था, Art and Architecture / कला और स्थापत्य, India after Independence / स्वतंत्रता के बाद भारत, General Awareness (Legacy Paper 1) / सामान्य जागरूकता (पुराना पेपर 1)
  6. Paper - 1 | Unit - 5 | Mathematical Reasoning and Aptitude: Number System and Simplification / संख्या पद्धति एवं सरलीकरण, Algebra and Ages / बीजगणित एवं आयु, Percentage / प्रतिशतता, Ratio, Proportion, Partnership and Mixtures / अनुपात, समानुपात, साझेदारी एवं मिश्रण, Averages / औसत, Profit, Loss and Discount / लाभ, हानि एवं बट्टा, Simple and Compound Interest / साधारण एवं चक्रवृद्धि ब्याज, Time and Work / समय एवं कार्य, Speed, Time and Distance / चाल, समय एवं दूरी, Mensuration and Geometry / क्षेत्रमिति एवं ज्यामिति, Series / श्रेणी, Coding-Decoding / कूटलेखन-कूटवाचन, Relationships, Directions and Ranking / संबंध, दिशा एवं क्रम, Arrangements, Puzzles, Clock and Calendar / व्यवस्था, पहेलियाँ, घड़ी एवं कैलेंडर, Mathematical Foundations, जीवनी, Philosophy of the Constitution / संविधान का दर्शन, Analysis, Indian Polity and Governance (Legacy Unit 10) / भारतीय राजव्यवस्था और शासन (पुराना यूनिट 10)
  7. Paper - 1 | Unit - 6 | Logical Reasoning: Structure of Arguments / युक्ति की संरचना, Terms, Definitions and Uses of Language / पद, परिभाषा एवं भाषा के उपयोग, Categorical Propositions / निरुपाधिक प्रतिज्ञप्तियाँ, Categorical Syllogism / निरुपाधिक न्यायवाक्य, Venn Diagrams / वेन आरेख, Deduction, Induction and Analogy / निगमन, आगमन एवं सादृश्यता, Fallacies / तर्कदोष, Indian Logic: Pramanas / भारतीय तर्कशास्त्र: प्रमाण, Anumana and Vyapti / अनुमान एवं व्याप्ति, Hetvabhasa / हेत्वाभास, Non Verbal Reasoning (Spatial Aptitude) (Spatial Reasoning) (Visual Reasoning), Aristotle / अरस्तू, Western Schools of Philosophy / पाश्चात्य दर्शन की विचारधाराएँ, Kant / कांट, Sadhana Paada, Vaiśeṣika / वैशेषिक
  8. Paper - 1 | Unit - 7 | Data Interpretation: Data and Graphical Representation / आँकड़े एवं आलेखीय निरूपण, Data Interpretation: Tables / आँकड़ा निर्वचन: तालिका
  9. Paper - 1 | Unit - 8 | Information and Communication Technology: Computer Fundamentals and Hardware, Memory, Storage and Data Representation, Computer Software, Office Applications, Multimedia and File Formats, Computer Networks, Internet and World Wide Web, Email and Online Communication, Cyber Security, Digital Ethics, IPR and Society, Cloud and Distributed Computing, Emerging Technologies, ICT in Education and Governance, Cyber Threats and Malware Attacks, Role of Internet and Web Services, Electronic Data Interchange and E-Commerce, Internet Fundamentals, Web Design & Development, Web Publishing & Hosting, Emerging Technologies, Terms & Abbreviations, Society, Law & Ethics, Keyboard Shortcuts, Website, Browser & Services, Protection of Geographical Indications / भौगोलिक संकेतों का संरक्षण
  10. Paper - 1 | Unit - 9 | People Development and Environment: Ecosystems and Biodiversity / पारितंत्र और जैव विविधता, Human and Environment Interaction / मानव और पर्यावरण की अंतःक्रिया, Development and Environment / विकास और पर्यावरण, Natural Resources – Forests, Water and Soil / प्राकृतिक संसाधन: वन, जल और मृदा, Energy Resources / ऊर्जा संसाधन, Air Pollution / वायु प्रदूषण, Water Pollution and Treatment / जल प्रदूषण और उपचार, Noise Pollution / ध्वनि प्रदूषण, Waste Management / अपशिष्ट प्रबंधन, Environmental Health and Toxicology / पर्यावरणीय स्वास्थ्य और विषविज्ञान, Climate Change and Global Environmental Issues / जलवायु परिवर्तन और वैश्विक पर्यावरणीय समस्याएँ, Natural Hazards and Disaster Management / प्राकृतिक प्रकोप और आपदा प्रबंधन, Indian Environmental Laws and Policies / भारतीय पर्यावरण कानून और नीतियाँ, International Environmental Agreements / अंतर्राष्ट्रीय पर्यावरण समझौते, Environment / पर्यावरण, Geography / भूगोल, Factory and Industry Systems / कारखाना और उद्योग व्यवस्थाएँ, Earth Materials, Surface Features and Processes, Communicable Diseases and Immunization / संक्रामक रोग और प्रतिरक्षण
  11. Paper - 1 | Unit - 10 | Higher Education System: Education in Ancient India / प्राचीन भारत में शिक्षा, Ancient Centres of Higher Learning / प्राचीन भारत के उच्च शिक्षा केंद्र, Colonial Education Policies and Commissions / औपनिवेशिक काल की शिक्षा नीतियाँ और आयोग, Colonial Colleges, Universities and Reformers / औपनिवेशिक कॉलेज, विश्वविद्यालय व सुधारक, Education Commissions and Policies after Independence / स्वतंत्र भारत में शिक्षा आयोग और नीतियाँ, National Education Policy 2020 / राष्ट्रीय शिक्षा नीति 2020, Types of Higher Education Institutions / उच्च शिक्षा संस्थानों के प्रकार, Learning Programmes and Academic Credits / अधिगम कार्यक्रम और अकादमिक क्रेडिट, Professional, Technical and Skill-Based Education / व्यावसायिक, तकनीकी और कौशल आधारित शिक्षा, Value Education and Environmental Education / मूल्य शिक्षा और पर्यावरण शिक्षा, Governance, Administration and Financing / शासन, प्रशासन और वित्तपोषण, Regulation, Accreditation and Ranking / उच्च शिक्षा का नियमन, प्रत्यायन और रैंकिंग
  12. Paper 2 | Unit 1 | Discrete Structures and Optimization: Propositional and Predicate Logic, Set Theory, Relations, Functions, Permutation, Combination and Probability / क्रमचय, संचय एवं प्रायिकता, Graph Theory, Group Theory, Digital Systems & Boolean Basics, Boolean Expression, Boolean Minimization, Optimization, Markov Chains
  13. Paper 2 | Unit 2 | Computer System Architecture: Logic Gates & Hardware, Combinational Circuit, Sequential Circuits, Number System, Number Representation, Floating Point Rep, Basics of COA, Register Transfer and Microoperations, Programming the Basic Computer, Instr Formats & Modes, Control Unit Design, Pipelining, Input Output Organisation, Cache Memory Organization, Multiprocessors
  14. Paper 2 | Unit 3 | Programming Languages and Computer Graphics: Language Design, C Fundamentals, Control Flow, Functions, Arrays & Pointers, Storage Classes, Structures & Enums, DMA, Macros, Scoping & File Handling, HTML Basics, XML, JavaScript-Basics, Java, Basics of Computer Graphics, 2-D Geometrical Transforms and Viewing, 3-D Object Representation, Geometric Transformations and Viewing, OOPS with C++, Java Fundamentals, Array, String, IO & References, OOPS with Java
  15. Paper 2 | Unit 4 | Database Management Systems: Basics of DBMS, ER Diagram, Relational Model & Functional Dependencies, Keys & Integrity Constraints, Normalization (1NF - BCNF), Decomposition Properties & 4NF, File Organization & Indexing, Relational Algebra, SQL, Relational Calculus, Transaction Management, Concurrency Control, Database Recovery, ORDBMS, Database Security & Authorization, Query Processing & Optimization, Enhanced Data Models, Data Warehousing & Mining, Big Data Systems, NoSQL
  16. Paper 2 | Unit 5 | System Software and Operating System: Introduction to OS, Process Management, CPU Scheduling, Process Synchronization, Threads & Process Creation, Deadlock, Memory Management, Virtual Memory, Disc Scheduling, File Management, Windows OS, Linux OS, Security, Distributed Systems, Virtual Machines
  17. Paper 2 | Unit 6 | Software Engineering: Fundamentals of Software Engineering, Software Requirements and Quality Assurance, Software Design, Estimation and Metrics, Software Testing, Software Maintenance and Configuration Management
  18. Paper 2 | Unit 7 | Data Structures and Algorithms: Introduction to DS, Array, Stack, Queue, Linked List, Tree, Graphs, Hashing, Algorithm Analysis, Time Complexity Analysis, Sorting Algorithms, Greedy Algorithms, Dynamic Programming, Minimum Spanning Trees, Shortest Path Algos, Advanced Algorithms
  19. Paper 2 | Unit 8 | Theory of Computation and Compilers: Introduction to TOC, Deterministic FA (DFA), Non-Deterministic FA, Regular Expressions, Grammar, Regular Language Properties, Moore & Mealy Machines, Pushdown Automata & CFG, Turing Machines, Complexity Theory, Intro to Compilers, Lexical Analysis, Grammar & CFG, Syntax Analysis: Top-Down, Syntax Analysis: Bottom-Up, Semantic Analysis & SDT, Intermediate Code Gen, Code Optimization, Run Time Environment
  20. Paper 2 | Unit 9 | Data Communication and Computer Networks: Introduction to CN, Data Communication, DLL: Access Control, DLL: Flow Control, DLL: Error Control, DLL: Framing, Data Link Layer - Ethernet, Net Layer: IPv4 & Proto, Net Layer: IP Addressing, Net Layer:Routing Protocol, Transport Layer Services, TL: Congestion & UDP, Application Layer, Hardware basics, Network Security, Cloud Computing, Mobile Computing
  21. Paper 2 | Unit 10 | Artificial Intelligence: Approaches to AI, Search Algorithms, Game Playing, Knowledge Representation, Planning, Multi Agent Systems, Fuzzy Sets, Natural Language Processing, Artificial Neural Networks, Genetic Algorithms
  22. Full Notes: Paper 1 Full Notes
  23. Paper 1 | Full Mock Tests:
  24. Paper 2 | Full Mock Tests:
  25. Paper 1 | Previous Year Papers:
  26. Paper 2 | Previous Year Papers:
AI summary & chapters

AI Summary

An AI-generated summary of this video lecture.

The video presents a structured explanation of training discrete Hopfield networks (DHNs) using Hebbian learning rules for both binary and bipolar input patterns. In the initial segment (00:00-02:00), Case 1 introduces binary patterns with the weight matrix formula Σ[2s_i(p) - 1][2s_j(p) - 1] for i ≠ j, where binary values are transformed into bipolar form (-1 or +1) to enable storage of multiple patterns. The on-screen text explicitly displays this equation, reinforcing the mathematical foundation. Transitioning to Case 2 (02:00-05:00), the focus shifts to bipolar input patterns, where the weight matrix simplifies to Σs_i(p)s_j(p) for i ≠ j, with the formula clearly shown on screen. The instructor emphasizes that self-connections are excluded by setting ω_ii = 0, a detail reiterated in the subsequent discussion. The testing algorithm is then introduced: weights are initialized via Hebbian learning, initial activations y_i are set equal to external input vectors x_i, net inputs are computed as y_i = x_i + Σw_ij * y_j, and outputs are determined using an activation function (y_i = 1 if net input > θ_i, else -1). This process is repeated until activations converge. In the final segment (10:00-11:36), a concrete example is worked through using the input vector X = [1 1 1 -1], demonstrating how to construct the symmetric weight matrix W via the outer product X^T * X, with diagonal elements set to zero. The resulting matrix W = [[0, 1, 1, -1], [1, 0, 1, -1], [1, 1, 0, -1], [-1, -1, -1, 0]] is explicitly displayed, illustrating the network's structure before testing with asynchronous updates. The progression moves from theoretical formulation to practical application, grounding the abstract rules in a tangible example.

Chapters

  1. 0:00 – 2:00 00:00-02:00

    The video introduces the training of a discrete Hopfield network by presenting two cases for storing patterns: binary and bipolar input patterns. For Case 1, it derives the weight matrix formula for binary inputs as Σ[2s_i(p) - 1][2s_j(p) - 1] for i ≠ j, where s(p) represents binary patterns. For Case 2, it presents the weight matrix formula for bipolar inputs as Σ[s_i(p)][s_j(p)] for i ≠ j, emphasizing the difference in weight calculation based on input type. The content transitions to a testing algorithm that includes initialization of weights and iterative steps for activation consolidation, with Step 1 specifying weight initialization and Step 2 initiating the process if activations are not consolidated. The instructor uses handwritten annotations to highlight key terms like 'Weight Matrix' and employs arrows to indicate the flow from stored patterns to weight calculation, reinforcing the derivation process.

  2. 2:00 – 5:00 02:00-05:00

    The video segment explains the testing algorithm for a Hopfield neural network, focusing on bipolar input patterns. The instructor presents Step 4 of the algorithm: initializing network activation to match the external input vector X, with the formula y_i = x_i for i = 1 to n. This is followed by Step 6, where the net input y_i is calculated using the formula y_i = x_i + Σ w_ij * y_j, incorporating both external input and weighted feedback from other units. Step 7 applies an activation function to determine output, defined as y_i = {1 if y_i > θ_i, -1 otherwise}, ensuring binary output based on threshold comparison. Step 8 involves broadcasting the computed y_i to all other units in the network, enabling distributed computation. Step 9 tests for convergence by checking if activations have stabilized across all units. The weight matrix is initialized using the Hebbian learning rule, with w_ij = Σ s_i(p) * s_j(p) for i ≠ j and ω_ii = 0, derived from stored patterns. The instructor emphasizes the iterative nature of this process and how it enables pattern retrieval through network dynamics.

  3. 5:00 – 10:00 05:00-10:00

    The video demonstrates the training of a discrete Hopfield network using an auto-associative approach with binary input vectors. The process begins by initializing the weight matrix W using the formula W = Σ[s^T(p)][s(p)] for i ≠ j, where s(p) is the input vector [1 1 1 -1]. The instructor shows that diagonal elements are set to zero, reflecting the absence of self-connections in the network. The weight matrix is computed as the outer product of the input vector with itself, resulting in a symmetric matrix: W = [[0, 1, 1, -1], [1, 0, 1, -1], [1, 1, 0, -1], [-1, -1, -1, 0]]. The explanation emphasizes that the network is symmetric and lacks self-connections. This weight matrix is then prepared for testing, where asynchronous updates will be used to reconstruct the input pattern. The instructor walks through each step of weight initialization, highlighting key formulas and matrix operations on the board.

  4. 10:00 – 11:36 10:00-11:36

    In the final segment of the lesson, the training process for a discrete Hopfield network is completed with an example using binary input vectors. The instructor demonstrates initializing the weight matrix by computing the outer product of the input vector X = [1 1 1 -1] with itself, resulting in a symmetric matrix W. Diagonal elements are explicitly set to zero because self-connections are not allowed in the network, as indicated by the on-screen notation w_ii = 0. The resulting weight matrix is written out as W = [0 1 1 -1; 1 0 1 -1; 1 1 0 -1; -1 -1 -1 0], with the symmetry condition w_ij = w_ji emphasized. The lesson transitions to testing, where asynchronous updates are introduced as the method for evaluating network convergence when some entries in the input vector are missing. The on-screen text labels this as "Step 9: Test the network for conjunction," and the instructor prepares to simulate the update process, indicating that the trained weights will be used in a dynamic state evolution. The handwritten annotations and formula notation reinforce the mathematical steps, ensuring clarity on how the network's connectivity is derived from the input pattern.

This lesson segment provides a structured explanation of training discrete Hopfield networks (DHNs) using Hebbian learning rules, covering both binary and bipolar input patterns. It begins with theoretical derivations of weight matrix formulas—Σ[2s_i(p) - 1][2s_j(p) - 1] for binary inputs and Σs_i(p)s_j(p) for bipolar inputs—emphasizing the transformation of binary values to bipolar form and the exclusion of self-connections via ω_ii = 0. The teaching progression moves to a testing algorithm that includes weight initialization, activation setup (y_i = x_i), net input computation (y_i = x_i + Σw_ij * y_j), and activation function application (y_i = 1 if net input > θ_i, else -1), repeated until convergence. A concrete example using the vector X = [1 1 1 -1] demonstrates weight matrix construction via outer product W = X^T * X, with diagonal elements set to zero. The resulting symmetric matrix W = [[0, 1, 1, -1], [1, 0, 1, -1], [1, 1, 0, -1], [-1, -1, -1, 0]] is explicitly shown.

Loading lesson…