Page 115 - Touhpad Ai

∑ ∑ Speaker labelling: Identifies different speakers in group conversations and tags their parts in the
transcript.
∑ ∑ Acoustic training: Helps the system adjust to variations in voice pitch, tone, loudness, and background
noise.
∑ ∑ Profanity filtering: Detects and removes inappropriate words or phrases.
3. What is language translation? Also, lists some examples of language translation.
Ans. Language translation is one of the most impactful applications of NLP, allowing for the automatic
conversion of text or speech from one language to another. This process is essential in today's globalised
world, where communication across language barriers is critical for business, diplomacy, education, and
more. NLP-powered translation systems, which once relied on rule-based and statistical approaches, have
now evolved to use advanced deep learning models, particularly neural machine translation (NMT), which
significantly enhances the quality and accuracy of translations.
Some real-world examples of language translation are as follows:
∑ Google Translate: One of the most popular and widely used translation tools, Google Translate, leverages
NMT to provide instant translation between over 100 languages. By continuously improving its neural
translation models and incorporating massive datasets, Google Translate provides not only word-for-
word translations but also contextually aware results.

∑ DeepL Translator: DeepL is another cutting-edge translation service that uses advanced neural network
architectures to outperform other translation tools in terms of accuracy and fluency.
∑ Microsoft Translator: Microsoft also offers a neural machine translation service, which supports text,
speech, and even image translation.

4. Explain any three types of Internet searches.
Ans. Some types of internet searches are as follows:
∑ Navigation searches: Navigation searches help users find a specific, known website or platform, often
related to popular brands or domains such as LinkedIn or YouTube. NLP identifies brand names, proper
nouns, and phrases indicating a user’s intent to navigate directly. Through entity recognition and query
classification, the system accurately directs users, prioritising official site links over generic content. For
instance, searches like “YouTube login” or “LinkedIn profile” will lead users straight to those platforms.
∑ Informational searches: Informational searches are made by users looking for knowledge or answers,
without the intent to make a transaction. NLP enhances these searches by applying semantic parsing,
question-answering models, and entity extraction to interpret the user's intent and deliver precise,
contextually relevant information. For example, queries like “What is artificial intelligence?” or “Symptoms
of COVID” are understood as requests for factual knowledge rather than commercial content.
∑ Transactional searches: Transactional searches indicate a user's intent to perform specific actions, such
as purchasing, downloading, or signing up for services. NLP detects action verbs like “buy,” “order,” or
“download” in conjunction with product or service names. This enables the search engine to provide
relevant transactional results, such as e-commerce pages or download links. Examples include searches
like “Buy iPhone 15” or “Download Netflix app,” which lead users directly to purchase or download options.

5. Lists the various use of AI in smartphones.
Ans. AI is deeply embedded in modern smartphones, enhancing everything from user experience to device
performance:
∑ Voice assistants: Smartphones like iPhones (Siri), Android devices (Google Assistant), and Samsung
devices (Bixby) utilise AI to understand voice commands, perform tasks (set reminders, send messages,
etc.), and provide information (weather, directions, etc.).
∑ Face recognition: AI is used for biometric authentication, allowing users to unlock their phones, make
payments, or access apps using face recognition (such as Face ID in Apple devices).

Introduction and State of Art of AI, Natural Language Processing (NLP), and Potential use of AI 113