Network Terms to Learn - Kalai Selvi Arivalagan (best love novels of all time .TXT) 📗

erased by accident, all while remaining as close as possible to the metal of the device.

Firmware comes in various complexities and can be found in simple devices, like keyboards and hard drives, to more complex ones, like graphics cards and Basic Input/Output System (BIOS).  In Android operating systems, the firmware is different depending on the manufacture, that is, it is the operating software is device-specific.  When a device is powered on, firmware is the first part to run and starts sending instructions to the device's processor to execute. If the device is as simple as a keyboard, the firmware does not stop working as there is no software to replace it. However, in more complex devices, such as PCs, laptops, and tablets, there are often multiple firmware sets that interact to achieve a common goal; load the operating system.

Regardless of the type of device, firmware can only work with a basic or low level, binary language known as machine language. While the firmware's code could be written in a high level language for ease and versatility, it needs to be translated into a low level language before getting etched into the device. The same firmware cannot run on processors it was not designed for, as different processors can only identify certain instructions. If a device’s firmware were to get corrupted—during an update, for example—it cannot be fixed, as there is no way to communicate with the machine to install a replacement.

 

Hyperautomation

Hyperautomation is a strategic approach to scaling automation within an enterprise. The strategy involves identifying what tasks to automate and choosing the most appropriate automation tools for each task.  An important goal of hyperautomation is to optimize ways that robotic process automation (RPA) can be used to improve productivity.  Gartner and some big companies in the tech space are promoting hyperautomation as the next wave of automation or “automation 2.0.” 

 

Hyperparameter

A hyperparameter is a machine learning parameter whose value is chosen before a learning algorithm is trained.

Hyperparameters should not be confused with parameters . In machine learning, the label parameter is used to identify variables whose values are learned during training. The prefix hyper is used to identify higher-level parameters that control the learning process.

Every variable that an AI engineer or ML engineer chooses before model training begins can be referred to as a hyperparameter -- as long as the value of the variable remains the same when training ends.

It’s important to choose the right hyperparameters before training begins because this type of variable has a direct impact on the performance of the resulting machine learning model. Examples of hyperparameters in machine learning include:

Model architecture Learning rate Number of epochs Number of branches in a decision tree Number of clusters in a clustering algorithm

 

Robotic Process Automation (RPA)

 

Robotic process automation (RPA) is a technology that uses software agents (bots) to carry out routine clerical tasks without human assistance. RPA is useful for automating business processes that are rules-based and repetitive.

RPA bots can follow a workflow that encompasses multiple steps across multiple applications. Unlike traditional automation projects that require extensive developer help, RPA projects simply use an organization's existing applications.

Essentially, RPA can be thought of as a more sophisticated version of macros. Initially, the technology requires a human to record themselves carrying out a specific business process. This creates a script that a bot uses to replicate workflow.

RPA is often used for data preprocessing tasks, including data entry, data reconciliation and spreadsheet manipulation. Additional uses at the enterprise-level include data analytics, data reporting and event-driven customer outreach.

Popular commercial off-the-shelf (COTs) RPA tools include Blue Prism, Automation Anywhere and UiPath.

 

 

 

Term of the day - 39

 

Synthetic Data

Synthetic data is input that is generated mathematically from a statistical model. Synthetic data plays an important role in finance, healthcare and artificial intelligence (AI) when it is used to protect personally identifiable information (PII) in raw data and fabricate massive amounts of new data to train machine learning (ML) algorithms.

Synthetic data is created by executing sequential statistical regression models against each variable in a real-world data source. Any new data collected from the regression models will statistically have the same properties as the originating data, but its values will not correspond to a specific record, person or device.

Synthetic data provides data scientists and analysts with quick access to additional data and frees them from having to worry about compliance. Its varied uses include:

Machine learning (ML) -- synthetic data can be used to quickly create additional data that statistically resembles the originating raw data.
Analytics -- synthetic data can be used to build large datasets by extrapolating information from relatively small datasets.
Compliance -- synthetic data can be used to provide data privacy by de-coupling the information a record contains from its originating source.
Information security -- synthetic data can be used to populate honeypots with fabricated data that's realistic enough to attract attackers.
Software development -- synthetic data can be used in quality assurance (QA) to test code changes in a sandbox environment.

 

Transport Layer

 

The transport layer is the fourth layer in the open systems interconnection (OSI) network model.  The OSI model divides the tasks involved with moving information between networked computers into seven smaller, more manageable task groups. Each of the seven OSI layers is assigned a task or group of tasks.

The transport layer's tasks include error correction as well as segmenting and desegmenting data before and after it's transported across the network. This layer is also responsible for flow control and making sure that segmented data is delivered over the network in the correct sequence.  Layer 4 (the transport layer) uses the transmission control protocol (TCP) & user data protocol (UDP) to carry out its tasks.

 

Overfitting

Overfitting is a condition that occurs when a machine learning or deep neural network model performs significantly better for training data than it does for new data.  Overfitting is the result of an ML model placing importance on relatively unimportant information in the training data. When an ML model has been overfit, it can't make accurate predictions about new data because it can't distinguish extraneous (noisey) data from essential data that forms a pattern.

For example, if a computer vision (CV) program's task is to capture license plates, but the training data only contains images of cars and trucks, the learning model might overfit and conclude that having four wheels is a distinguishing characteristic of license plates. When this happens, the CV programming is likely to do a good job capturing license plates on vans, but fail to capture license plates on motorcycles.

The most common causes of overfitting include the following:

The data used to train the model is dirty and contains large amounts of noise.
The model has a high variance with data points that are very spread out from the statistical mean and from each other.
The size of the training dataset is too small.
The model was created by using a subset of data that does not accurately represent the entire data set.

Computer Vision

 

Computer vision (CV) is the subcategory of artificial intelligence (AI) that focuses on building and using digital systems to process, analyze and interpret visual data. The goal of computer vision is to enable computing devices to correctly identify an object or person in a digital image and take appropriate action.

Computer vision uses convolutional neural networks (CNNs) to processes visual data at the pixel level and deep learning recurrent neural network (RNNs) to understand how one pixel relates to another.

Uses for computer vision include:

Biometric access management -- CV plays an important role in both facial and iris recognition.
Industrial robots and self-driving cars -- CV allows robots and autonomous vehicles to avoid collisions and navigate safely.
Digital diagnostics -- CV can be used in tandem with other types of artificial intelligence programming to automate the analysis of X-rays and MRIs.
Augmented reality -- CV allows mixed reality programming to know where a virtual object should be placed.

 

Personally Identifiable Information (PII)

Personal Identifiable Information (PII) is a label used to describe data that directly or indirectly identifies a specific individual.

Examples of PII include names, addresses, biometrics and alphanumeric account numbers.

Name -- includes full names, maiden names, mother‘s maiden names, nicknames and aliases.
Address -- includes street addresses, email addresses, IP addresses and MAC addresses.
Biometrics -- includes photographs, x-rays and other types of bio-based data such as fingerprints.
Alphanumeric account numbers -- includes telephone numbers, driver‘s license numbers, taxpayer IDs, patient IDs, vehicle registration numbers and credit card numbers.
In many parts of the world, personally identifiable data has to be collected, stored and destroyed in accordance with compliance rules and regulations. Because non-PII can easily become PII if additional information is made publicly available, this type of data should be periodically reviewed to determine whether its IT risk management level has changed.

Risk impact levels (low, medium, high) for PII are subjective and based on the potential harm that inappropriate access, use or disclosure of the personally identifiable information would cause. The likelihood of risk is greatly reduced if an organization minimizes the amount of PII it collects, stores and shares.

 

Logistic Regression

Logistic regression is a supervised learning algorithm used in machine learning to predict the probability of a binary outcome. A binary outcome is limited to one of two possible outcomes. Examples include yes/no, 0/1 and true/false.  Logical regression is used predictive modeling to analyze large datasets in which one or more independent variables can determine an outcome. The outcome is expressed as a dichotomous variable that has one of two possible outcomes.  Essentially, logistic regression works by estimating the mathematical probability that an instance belongs to a specified class -- or not.

 

DevRel (DeveloperRelations)

 

Developer Relations (DevRel) is a strategic approach to improving business-to-business (B2B) communications between a software company’s internal programmers and the external programmers who will be using the company’s open APIs.  Just as DevOps strives to improve communication between an organization’s developer and operations teams, DevRel seeks to improve communication between proprietary and open-source developer communities.

Typically, a DevRel team is made up of employees from the software company's product, engineering and marketing teams. In addition to serving as developer evangelists, team members are responsible for gathering metrics that can be used to quantify user engagement and help the software company's internal developers understand their end users' pain points.

 

Artificial Intelligence Engineer

An artificial intelligence engineer is someone whose job is to identify the right approach to using AI to solve a specific business problem. In the enterprise, AI engineers typically work closely with machine learning engineers to develop and deploy learning algorithms that can use historical and real-time data to predict future events. In smaller companies, the same person (AI/ML engineer) may be responsible for both AI strategy and implementation.

AI engineers need a strong background in math and statistics. Ideally, they are also familiar with Python and R, as well as their most commonly-used libraries and packages. Basic responsibilities for an AI engineer include the following:

Establish business objectives for implementing AI. Brainstorm with other IT team members to explore how machine learning concepts can be used solve specific business problems. Develop proof of concepts (POCs). Identify obstacles that could potentially put an AI-driven project at risk and research workarounds. Create metrics to measure an AI project’s return on investment (ROI). Promote best practices for data wrangling, data processing and project documentation

 

Machine Intelligence

 

Machine intelligence is an umbrella term that's used to describe the accuracy of a machine learning (ML), deep learning (DL) or classical algorithm output.

In business-to-business (B2B) marketing, the term is also being used to describe a growing market