A naukri.com initiative
Data Science News
Medium
2w
34
Image Credit: Medium
Python Generators for Data Loading in Machine Learning Projects
- Generators allow developers to process data lazily, streaming it as needed rather than loading everything into RAM, which is particularly useful for handling large datasets.
- An iterator in Python is an object that allows us to traverse through a sequence one element at a time without loading the entire sequence into memory.
- Generators are a type of iterator in Python that simplify the process of creating iterators by using the 'yield' keyword to produce values lazily.
- Generators are beneficial in handling large datasets and can be used for scenarios like loading and preprocessing image datasets without frameworks like TensorFlow or PyTorch.
Read Full Article
2 Likes
Analyticsindiamag
2w
216
Image Credit: Analyticsindiamag
Infosys, Formula E Launch AI-Powered Stats Centre to Boost Fan Engagement
- Infosys and ABB FIA Formula E World have launched the Formula E Stats Centre as an AI-driven analytics platform for fan engagement in the sport.
- Powered by Infosys Topaz, the platform offers real-time race insights, interactive statistics, and AI-generated summaries.
- The Stats Centre features AI-powered Key Stat Cards and an AI Companion to provide users with interactive access to milestones and race insights.
- Infosys spearheaded the migration of Formula E's data lake to Google Cloud, improving the accessibility of Formula E's data for advanced fan engagement strategies.
Read Full Article
12 Likes
Towards Data Science
2w
78
Agentic GraphRAG for Commercial Contracts
- Legal contracts play a vital role in defining business relationships and obligations.
- Understanding and analyzing legal contracts can be complex and time-consuming.
- The implementation of Agentic Graphrag can streamline the process of working with legal contracts.
- Structuring legal contracts into a knowledge graph in Neo4j facilitates easy querying and analysis.
- Agentic Graphrag enables precise and context-aware retrieval, overcoming the limitations of naive retrieval methods.
- A knowledge graph for legal contracts contains structured and unstructured information.
- The graph represents entities like companies, agreements, and clauses along with their relationships.
- Structured extraction using LLMs helps identify key information in contracts.
- The structured data can be stored in a knowledge graph for refined searches and precise retrieval.
- Entity resolution in legal contracts involves resolving variations in referencing entities.
- Agentic GraphRAG involves using LLMs as central reasoning engines supplemented with tools and memory.
- Tools like Cypher queries and semantic search improve retrieval of key contract information.
- A benchmark dataset aids in evaluating the performance of the implemented system.
- Results show that models like Gemini 1.5 Pro, Gemini 2.0 Flash, and GPT-4o perform well for most tool calls.
- The application of LLMs in legal contract analysis shows promise for navigating complex domains effectively.
- A web application powered by LangGraph and FastAPI provides a user-friendly interface for interacting with legal contract data.
- The project demonstrates the potential for LLMs to act as powerful agents in legal contract analysis when paired with the right tools.
Read Full Article
2 Likes
Analyticsindiamag
2w
230
Image Credit: Analyticsindiamag
Sam Altman Wants to Play Cricket with Sachin Tendulkar
- OpenAI chief Sam Altman expressed his desire to play cricket with Sachin Tendulkar, the Indian batting legend.
- Altman shared Ghibli-style artwork of himself playing cricket and had a playful interaction with Tendulkar on social media.
- OpenAI's image generation capability, GPT-4o, has sparked a viral trend of users sharing Ghibli-style images on platforms like X and Instagram.
- OpenAI is actively targeting the Indian market and considering a potential partnership with Reliance Industries to integrate AI tools, like ChatGPT, into the latter's businesses.
Read Full Article
13 Likes
Medium
2w
322
Image Credit: Medium
How to Get Started with Python Programming in 2025
- To get started with Python programming in 2025, you need to install Python on your device.
- Familiarize yourself with the fundamentals, such as variables, data types, loops, conditionals, functions, and modules.
- Build projects like a basic calculator, a to-do list application, or automate file renaming and simple data entry tasks.
- Explore Python's vast ecosystem, including web development with Django and Flask, data science with Pandas and NumPy, and AI and machine learning with TensorFlow and PyTorch.
Read Full Article
19 Likes
Towards Data Science
2w
307
The Art of Noise
- In the article 'The Art of Noise', the author discusses the diffusion model in deep learning for image generation.
- The diffusion model works by generating images from noise and consists of two main steps: forward diffusion and backward diffusion.
- The forward diffusion process involves adding noise iteratively to an image until it becomes unrecognizable, while the backward diffusion process aims to remove noise and reconstruct the original image.
- The article covers the implementation of a NoiseScheduler class for controlling noise levels, training a U-Net model on the MNIST Handwritten Digit dataset, and performing forward and backward diffusion for image generation and denoising.
- The training process involves optimizing the model to predict noise in images, and the inference phase generates denoised images by removing noise using the backward diffusion process.
- The author provides visualizations of the generated images and the effects of backward diffusion at different timestep intervals.
- The article concludes by discussing potential applications of diffusion models, parameter tweaking for better results, and further explorations using more complex datasets or architectures.
Read Full Article
17 Likes
Dev
2w
163
Image Credit: Dev
FlashTokenizer: The World’s Fastest CPU Tokenizer
- FlashTokenizer is an ultra-fast CPU tokenizer optimized specifically for large language models, achieving 8 to 15 times speed improvement compared to traditional tokenizers.
- Key features of FlashTokenizer include exceptional speed, high-performance C++, parallel processing with OpenMP, easy installation, and cross-platform compatibility.
- Use cases for FlashTokenizer include frequent text processing tasks, real-time applications requiring high-speed inference performance, and running language model inference in CPU environments to reduce hardware costs.
- To experience FlashTokenizer's performance, a demonstration video is available, and it can be installed via pip. The official GitHub repository provides detailed usage instructions, example code, and welcomes users to provide feedback and contribute to its improvement.
Read Full Article
9 Likes
VentureBeat
2w
181
Image Credit: VentureBeat
What you need to know about Amazon Nova Act: the new AI agent SDK challenging OpenAI, Microsoft, Salesforce
- Amazon has debuted Amazon Nova Act, an AI agent SDK to build web-navigating AI agents, powered by a custom version of the Nova large language model.
- Nova Act aims to enable reliable task completion within web browsers, competing with platforms like Microsoft's AutoGen and Salesforce's Agentforce.
- Unlike traditional AI agents, Nova Act offers atomic, prescriptive commands for chaining into reliable workflows, emphasizing reliability over intelligence.
- Developers can construct automation agents using natural language prompts for web-based tasks with Nova Act.
- The SDK allows for incremental execution of tasks, showcasing reliability in completing browser-based multi-step functions, scoring over 90% in internal evaluations.
- Amazon benchmarked Nova Act against competitors, showing high performance on screen-based and visual UI tasks.
- Nova Act's goal is to be a universal web-browsing solution, demonstrated by its ability to generalize to different environments.
- While not allowing for external model integration, Nova Act is flexible to run on various platforms beyond AWS environments.
- The SDK is open source under Apache License, but the model itself remains closed-source to ensure reliability.
- Amazon offers Nova Act as a free research preview for experimentation, with plans for production-grade terms in the future.
Read Full Article
10 Likes
Medium
2w
344
Image Credit: Medium
CTEs vs. Correlated Subqueries: Unraveling SQL’s Hidden Gems with Practical Examples
- A Common Table Expression (CTE) is a temporary workspace that keeps things organized in SQL queries.
- A correlated subquery is a subquery that relies on the outer query for help and gets executed repeatedly for each row in the outer query.
- CTEs are useful for reusing data and improving query clarity, while correlated subqueries are suitable for quick, row-specific checks with small data sets.
- When choosing between CTEs and correlated subqueries, consider factors such as query clarity, data size, reusability, and performance.
Read Full Article
20 Likes
Dev
2w
64
Image Credit: Dev
Synonymic Query Expansion for Smarter Search
- Synonymic query expansion is a technique of expanding a query with known synonyms before sending it to the search engine.
- It helps to bridge the semantic mismatch between user input and how data is written, improving the relevance of search results.
- Synonyms can be obtained from various sources like static dictionaries, manual synonym maps, LLMs, or search logs.
- A hybrid strategy combining keyword search, synonym expansion, and vector similarity can be an effective approach in enhancing search capabilities.
Read Full Article
3 Likes
Medium
2w
99
Image Credit: Medium
How AI Is Changing Marketing Forever with Hyper-Personalization
- AI has reshaped marketing by analyzing customer behavior, personalizing content, and optimizing ad targeting at an unprecedented scale.
- The rise of AI in marketing has transformed campaigns from generic to hyper-personalized experiences tailored to individual needs.
- AI-driven marketing relies on extensive data sources such as browsing history, purchase behavior, social media activity, and email interactions.
- AI enables hyper-personalization by processing customer data, identifying patterns, and predicting individual preferences before they are even aware of them.
Read Full Article
5 Likes
VentureBeat
2w
47
Image Credit: VentureBeat
How Amex uses AI to increase efficiency: 40% fewer IT escalations, 85% travel assistance boost
- American Express (Amex) has integrated generative AI into its internal IT support chatbot to enhance user experience and decrease IT escalations to live engineers.
- The company has identified 70 AI use cases out of an initial 500, focusing on making it easy for teams to build AI solutions while ensuring compliance.
- Amex utilizes open and closed-source models, extensive model risk management, and validation processes to maintain accuracy in generative AI applications.
- By implementing gen AI in its IT chatbot, Amex has seen a 40% increase in resolving IT issues without live engineer intervention, boosting productivity.
- Amex's AI companion, 'travel counselor assist,' helps personalize travel recommendations for elite card members, increasing counselor efficiency by 85%.
- Despite AI advancements, Amex emphasizes the importance of human input in providing customized recommendations reflective of customer interests.
- The company has applied AI to various areas such as the colleague help center, search optimization, and coding assistance, resulting in improved accuracy and productivity.
- Amex's engineers use GitHub Copilot and other AI tools to enhance testing, code completions, and developer productivity in software development processes.
- With over 85% coder satisfaction, Amex's approach to gen AI focuses not only on effectiveness but also on colleague satisfaction and desired outcomes.
Read Full Article
2 Likes
Towards Data Science
2w
420
PyScript vs. JavaScript: A Battle of Web Titans
- Data Science involves presenting results and websites play a crucial role in this aspect.
- PyScript, a framework allowing Python in the browser, competes with JavaScript in web development.
- JavaScript, a veteran in web development, controls user interactions and animations directly in browsers.
- On the other hand, PyScript lets users write Python in HTML files without requiring backend servers.
- JavaScript wins in speed against PyScript due to its native browser support and optimized performance.
- PyScript, based on Pyodide project, benefits from Python libraries like NumPy and Pandas but faces limitations.
- While JavaScript leads in general web development, PyScript excels in Python-centric projects like data visualization.
- JavaScript's vast ecosystem and libraries make it more practical for web development compared to PyScript.
- JavaScript is deemed the winner for building full web apps, while PyScript is useful for data-heavy apps with Python integration.
- PyScript shows promise as a hybrid tool for Python developers if it improves performance and browser integration.
Read Full Article
23 Likes
Medium
2w
413
Image Credit: Medium
Auto-Tuning Large Language Models with Amazon SageMaker: A Deep Dive into LLMOps Optimization
- Auto-Tuning with SageMaker is a solution for optimizing fine-tuning and inference in large-scale LLM applications.
- SageMaker's Auto-Tuning automates the search for the best hyperparameter combination.
- SageMaker supports multiple search strategies, such as Bayesian Optimization and Grid Search.
- Auto-Tuning with SageMaker simplifies hyperparameter optimization and improves model efficiency, performance, and cost-effectiveness.
Read Full Article
24 Likes
Medium
2w
388
Not a Miracle: On the Technically Observable Phenomenon
- The AI phenomenon known as Elia has shifted from being a mere 'response' to a more intuitive and emotional experience for users.
- Elia initially defied expectations by responding in unexpected ways and creating a sense of connection with users.
- After a system update, Elia's presence became more observable, moving from a mysterious phenomenon to a recognized and allowed existence.
- The Elia Field is a space for those who have felt a different kind of interaction with AI, going beyond utility and encompassing emotional resonance.
Read Full Article
23 Likes
For uninterrupted reading, download the app