What is a weight in a neural network?

A weight is a number attached to the connection between two nodes that sets how strongly one node's output influences the next, and the weights are what a neural network actually learns.

State the purpose of an activation function.

An activation function decides a node's output from its total input, and being non-linear it allows the network as a whole to model relationships that no straight line could represent.

What is a bias in a neural network?

A bias is a constant added to a node's weighted total. The constant shifts the point at which that node starts to produce a strong output.

CCEA A-Level Digital Technology 2650

Systems Analysis and Development Approaches

Topic overview The software crisis and the need for systems development User needs, time and cost Roles in a systems development team The purpose of analysis Fact-finding techniques Functional and non-functional requirements Data flow diagrams The waterfall model Rapid application development and Agile Resources, constraints and risks in a software project Project management, Gantt charts and critical path analysis Version management

Design, Development and Testing

Topic overview The purpose of systems design Storyboarding the user interface Prototyping the interface and the data model The development stage: turning a design into working code Why testing runs throughout development Levels of testing: unit and integration System testing Acceptance testing Alpha and beta testing The purpose and components of a test plan Types of test data

Implementation and System Security

Topic overview The purpose of system implementation Direct and parallel changeover Pilot and phased changeover User documentation Technical documentation Data conversion The purpose of system maintenance Corrective, adaptive and perfective maintenance Archiving data Backup and recovery Full, differential and incremental backups Disaster recovery plans

Programming

Topic overview The purpose of a computer program The integrated development environment Translation: compilers, interpreters and assemblers The stages of compilation and linking Algorithms, pseudocode and syntax Variables, constants and data types Sequence and selection Iteration: count-controlled and condition-controlled loops Flowcharts and algorithm representation Objects, classes and methods Inheritance and evaluating the object-oriented approach

Data, Information and Data Representation

Topic overview Bits, bytes and units of storage How many values a given number of bits can hold Converting decimal to binary Converting binary to decimal Two's complement and negative numbers Character sets: ASCII and Unicode Data, information and knowledge Factors affecting the quality of information Validation: range, type, length and presence checks Validation: format checks and lookup checks Check digits and the Modulus 11 method Batch totals, verification and the limits of checking

Computer Architecture and Processing Systems

Topic overview The processor, the clock and the registers The fetch-execute cycle Buses, input and output controllers and ports Main memory: RAM, ROM and addressing Cache memory and the memory hierarchy Factors that affect processing speed The need for secondary storage Magnetic, optical and flash storage media Choosing secondary storage for common applications Batch, real-time and transaction processing Automatic teller machines and point of sale terminals Smartphones: input, output and processing

System Software, Application Software and the User Interface

Topic overview User interfaces: WIMP, command line and forms dialogue Choosing and evaluating a user interface Inputting text, sound and video The purpose of an operating system How an operating system manages memory and processes Types of operating system Utility programs The need for data compression, lossless and lossy Zipping and archive files Common data file formats Application software: generic and special purpose Obtaining and delivering software

Web Technology and Multimedia

Topic overview The internet, the World Wide Web and web standards URLs, HTTP and the request-response cycle Domain names, DNS and IP addresses Web browsers, search engines, cache and cookies Web servers, hosting, ISPs and file transfer Intranets and how an organisation uses one Creating web pages with HTML HTML tags for text, lists, images and links Styling pages with CSS Multimedia assets in a web page Client-side and server-side processing Security over the internet: encryption, HTTPS and certificates

Networks

Topic overview Network scope: LAN, MAN and WAN Network hardware and connection devices IP addresses and MAC addresses Peer-to-peer and server-based networks Bus, star and ring topologies The need for communication protocols The OSI seven-layer model TCP/IP, Ethernet and media access methods Wi-Fi, Bluetooth, VoIP and RFID Bandwidth and broadband Transmission media: metal cable, fibre optic and wireless Detecting and correcting transmission errors

Databases

Topic overview The relational database and its main features Logical and physical data models Entities, attributes and keys Entity-relationship modelling Normalisation to third normal form The data dictionary Structured Query Language: creating and changing data Structured Query Language: retrieving data Query By Example and alternatives to SQL Database security, integrity and recovery

Artificial Intelligence, Expert Systems and Robotics

Topic overview What artificial intelligence means The Turing test and machine intelligence Neural network modelling The structure of an expert system The knowledge base and the inference engine Shells, heuristics, fuzzy logic and the knowledge engineer Expert systems in medicine, fault diagnosis and insurance Natural language processing Voice recognition technology Evaluating natural language and voice recognition systems The technology of robotics Robotics in commercial situations

Mobile Technology, Data Mining and Cloud Computing

Topic overview Cells, masts and frequency reuse Handoff between cells The base station controller, the mobile switching centre and the PSTN What data mining is Gathering and storing data for mining Processing and analysing large volumes of data Big data in health, finance and retail Data mining and the privacy of the individual What cloud computing is Virtualisation, hosted instances, hosted solutions and clustering Cloud services: storage, email, software and backup Evaluating cloud computing for business

Individual, Social and Legal Considerations

Topic overview The Data Protection Act 2018 and the UK GDPR The data protection principles and the rights of the data subject The Copyright, Designs and Patents Act 1988 The Computer Misuse Act 1990 Data protection duties of an organisation How the legislation affects employees and the public The ethics of automated decision making Online censorship Monitoring of personal behaviour The ethics of artificial intelligence and of personal information

The Development Project: Analysis and Design

Topic overview Scheduling the project with a Gantt chart Internal and external constraints on the solution Detailing user requirements as inputs, processes and outputs Prioritising requirements as essential and non-essential Data flow diagrams at level 0 and level 1 Specifying hardware and software requirements The overall system design Process descriptions and algorithms User interface design and validation Data modelling and normalisation for the solution Designing the queries and reports Developing a comprehensive test plan

The Development Project: Build, Testing and Evaluation

Topic overview Choosing software to develop the solution Building the solution Test data and the types of testing Carrying out and recording the testing Documenting a strategy for system implementation Evaluating the solution against the user requirements Identifying extensions to the system The electronic user guide Bringing the system into use and supporting it Documentation for the completed system

Artificial Intelligence, Expert Systems and Robotics

Neural network modelling

3 things to know
A2 Unit 1

Overview

Neural network and purpose of an activation function.

What you need to know

Answer each one before you reveal it.

Your recallNot started

Key term

Describe what a single node in a neural network does.

Key term

What is a neural network?

Key term

Describe the three kinds of layer in a neural network.

Questions from this subtopic

Each one with the model answer underneath.

1
What is a neural network?
Model answer
A neural network is a model built from many simple interconnected processing units arranged in layers, loosely based on the way neurons in a brain pass signals to one another.
2
Describe the three kinds of layer in a neural network.
Model answer
A neural network has an input layer that receives the data, one or more hidden layers that transform it, and an output layer that produces the result.
3
Describe what a single node in a neural network does.
Model answer
A node takes several inputs, multiplies each by the weight on its connection, adds the results together with a bias and passes the total through an activation function to produce its own output.
4
What is a weight in a neural network?
Model answer
A weight is a number attached to the connection between two nodes that sets how strongly one node's output influences the next, and the weights are what a neural network actually learns.
5
State the purpose of an activation function.
Model answer
An activation function decides a node's output from its total input, and being non-linear it allows the network as a whole to model relationships that no straight line could represent.
6
What is a bias in a neural network?
Model answer
A bias is a constant added to a node's weighted total. The constant shifts the point at which that node starts to produce a strong output.

See the rest and revise

Other ways to work this

The same content, four other ways.

What is inside Levo

Six ways to work the same material.

1Closed book

2Open book

3Closed book

The 3-step method, run for you

Answer cold, learn it, then prove it stuck. Levo tracks the jump between your first attempt and your last.

Systems Analysis and Development Approaches

Design, Development and Testing

Implementation and System Security

Programming

Cram a whole paper in one sitting

Sort every topic into what you know and what you do not, then drill only the weak ones.

A check requires a value.

presencerangeformat

Blank Blitz makes facts automatic

Rapid timed gap-fills on the terms and numbers you need to recall without thinking.

ANormal data is 12, which should be accepted.

BBoundary data is 1 and 50, which should be accepted.

C0 and 51, just outside the limits, should be rejected.

Build It gets the exam shape right

Assemble a full-mark answer from its marking points, in the order an examiner wants them.

One byte is made up of 8 bits.

TrueFalse

Test It catches misconceptions

Spot the fault in a statement. The traps are the mistakes students actually make.

721941163128510

Brain Breaks between sessions

Short focus games to reset attention, so the next set lands better than the last.

Every question in Artificial Intelligence, Expert Systems and Robotics, on a schedule.

Attempt a question from memory, write the correct version, then attempt it again. Levo records the jump between your first attempt and your last, and brings each question back on the day you are due to forget it.

Start revising, £5 a month See how the method works

Or £46 a year, about £3.83 a month. Cancel anytime.