Vaishnov RG

Data Science Interview Question: Q. What's the difference between orthogonality and orthonormality in 2 vectors? Answer: Let's assume we have 2 vectors a and b. 1. Orthogonal - a and b are orthogonal if the dot product between them is zero. This means that there is no projection of one vector onto the other, implying that they are at a right angle (90 degrees) to each other. Mathematically - a⋅b=0 Orthogonality does not concern the magnitude of the vectors, only their angle. 2. Orthonormal - Two vectors are orthonormal if they are both orthogonal and each vector is of unit length. This means in addition to being orthogonal, each vector must have a magnitude (or norm) of 1. Mathematically, if a and b are orthonormal, they satisfy: a⋅b=0 and ∥a∥=1,∥b∥=1 Orthonormal vectors are a special case of orthogonal vectors where the additional condition of having unit norm is met.

125

4 Comments

Quantized embeddings offer significant advantages for NLP models, including reduced storage requirements, faster computations, lower energy consumption, and improved model interpretability. They are becoming increasingly important for enabling efficient and scalable NLP on resource-constrained devices and cloud-based applications. Read my latest article below

6

Question : Why does an AI product manager need to monitor #AI model❓ Once AI model is developed and deployed, the job doesnt end there. Model encodes pattern of the data it has seen during training phase. There may be several patterns or edge cases which are not part of training data. If model encounters any such patterns in testing or production phase, it will produce results which wont make sense without giving any warning about it. 👉 Data drift is a common challenge faced by AI-powered products. Over time, the distribution of input data can shift, leading to a mismatch between the model's training data and the real-world data it encounters. This can result in degraded model performance, unexpected outputs, and even complete model failure. 👉 The consequences of data drift can be severe, ranging from suboptimal user experiences to regulatory compliance issues. Product managers must be vigilant in detecting and addressing these challenges to ensure the continued success of their AI-powered offerings. 👉 Sudden changes in metrics can indicate input data drift. Additionally, monitoring the distribution of model outputs can reveal output drift, which may signal a need to retrain or fine-tune the model. Detecting data drift early is crucial. Product managers can leverage techniques such as A/B testing, anomaly detection, and continuous monitoring to identify and address these issues. Once detected, the product team can take appropriate actions, such as retraining the model, adjusting the data pipeline, or even considering a model architecture change. 💡 Quick Tip: Integrate model monitoring into your product roadmap and allocate dedicated resources to ensure the long-term success of your AI-powered offerings. Establish a comprehensive model monitoring dashboard to gain real-time visibility into the health and performance of your AI models. #artificialintelligence #productmanagement #datascience #productmanager

7

Understanding Temperature in Language Models 🌡️ Temperature is a pivotal parameter in large language models influencing the diversity and coherence of generated text through softmax-based sampling. Temperature is a parameter used in softmax-based sampling methods within language models. In essence, softmax converts logits (raw model outputs) into probabilities. Temperature scales these probabilities before sampling from them. The formula for temperature scaling is: Scaled logits = logits / T, where T is the temperature parameter. Higher temperatures lead to more uniform distributions over the vocabulary, encouraging the model to explore diverse word choices. Conversely, lower temperatures sharpen the distribution, favoring high-confidence predictions and reducing diversity in generated text. Diversity vs. Coherence: High Temperature: Increases the entropy of the probability distribution, encouraging the model to explore a wider range of word choices. This enhances creativity in generated text. Low Temperature: Decreases entropy, favoring high-confidence predictions and reducing randomness. This improves coherence and ensures more predictable outputs. Understanding temperature empowers researchers and developers to optimize language model outputs for specific applications, striking a balance between creativity and coherence. Let's delve deeper into the nuances of AI and NLP together! 💡 #NLP #AI #MachineLearning #DataScience #NaturalLanguageProcessing #LLMs Feel free to connect and share your experiences or ask any questions! 🙌

3

Week 10 of the "Ultimate Big Data Masters Program" provided an insightful journey into Apache Spark Optimizations and Performance Tuning - 1, provided by Sumit Mittal Sir. Here's a summary of Week 10 key learnings: ·       Pyspark Optimizations – Internals of groupBy ·       Normal Join Vs Broadcast Join ·       Different types of Joins ·       Partition Skew ·       Adaptive Query Execution (AQE) ·       Join Strategies ·       Optimizing Join of 2 large tables – Bucketing Thank you Sumit Mittal Sir and TrendyTech for all the support and guidance whenever needed. #Bigdata #Spark #Python #DAG #Join #dataengineer #AWS #Azure #ADLS #HDFS #Linux #SQL #athena #redshift

23

GPT4o for voice typing ------------------------------------- So far, accuracy with Sellinam, Google voice typing for #tamil is not 100%. After spelled, it requires edit. Tried #GPT4o to output the spoken words in Tamil. It is accurate and doesn't require any hard edit. கேட்பதும் நீ மறுப்பதும் நான் கெஞ்சுவதும் நீ இளகுவதும்... Disclaimer - Haven't tried complex sentences yet. Haven't used any other LLM for voice typing prior.

9

One more session alert by Vidvatta for people interested in prompt engineering. Topic - Efficient Prompt Engineering and Tuning for Language Models! 🗓️ Date: 26th May 2024 🕒 Time: 11 AM IST ⏳ Duration: 1 hour 30 mins 🔗 Register Here https://lnkd.in/gurGZmeV ✨ Join us in this webinar where you will have a unique experience of doing hands on together with mentor in this session and get live feedbacks when you are stuck. Are you interested in Large Language Models (LLMs) and their potential? Curious about how to leverage advanced prompting techniques for various tasks? Our upcoming webinar is perfect for you! ✨ What to Expect: A brief overview of LLMs and the importance of prompting. Prompting and Task Complexity- Classification of tasks: Easy, Medium, Hard. Advanced Prompting Techniques: * Zero-Shot and Few-Shot Learning: Definitions and applications. * Task Decomposition: Simplifying complex tasks. * Chain and Tree of Thought: Enhancing reasoning capabilities. Interactive Q&A Session: Open floor for attendee questions and sharing experiences. Future trends in LLMs and prompt engineering. Don't miss this opportunity to delve into the intricacies of LLMs and enhance your skills. Register now and be part of the conversation shaping the future of prompt engineering! In case of any queries contact support@vidvatta.in #Webinar #LLM #PromptEngineering #MachineLearning #ArtificialIntelligence #FutureTech #LearningCommunity

13

1 Comment

Connections, I have a 3GB file in.gz format. I am writing this file into an external table in adls location which is getting written in PARQUET format. The write job is taking around 27 mins for only 3 GB of data for only one date partition ; and it is over - partitioning the original file of 17 MB each and creating 200 part parquet files in adls. Its just reading and writing. No transformations in between. Workarounds that I have tried - 1) Incresed the number of executors per node. 2) Set the maxPartitionBytes spark parameter to a value targeting 5-6 partitions in adls. But its still taking 27 mins. Can anyone suggest how can I bring down the write time ?

71

31 Comments

Question : Sampling Techniques : What are they❓ As we saw recently in exit polls on Indian elections, choosing samples from a population can have a significant impact on predictions. #AI enthusiasts need to have a solid understanding of statistical sampling techniques. Statistical sampling is essential for AI product managers because it helps them: 1️⃣ Understand the Target Population: Accurate sampling ensures that the data used to train and evaluate AI models is representative of the target population, reducing the risk of biased or skewed results. 2️⃣ Optimize Data Collection: Sampling techniques can help AI product managers determine the appropriate sample size and sampling method, ensuring efficient and cost-effective data collection. 3️⃣ Validate Model Performance: Sampling is crucial for testing and validating AI models, as it allows for the creation of representative test sets that accurately reflect the real-world scenarios the model will encounter. AI enthusiasts should be aware of the different types of sampling methods, including: 🔺 Probabilistic Sampling Techniques: methods where randomisation is used instead of deliberate choice. 👉 Simple Random Sampling: Each member of the population has an equal chance of being selected. 👉Stratified Sampling: The population is divided into subgroups (strata), and samples are drawn from each stratum. 👉Cluster Sampling: The population is divided into clusters, and a random sample of clusters is selected. 👉Systematic Sampling: Every nth member of the population is selected, starting from a random point. 🔺 Non-Probabilistic Sampling Techniques: methods where some selection criteria exists 👉Convenience Sampling: Selecting participants based on their availability and accessibility. 👉Quota Sampling: Selecting participants to match predetermined proportions in the population. 👉Purposive Sampling: Selecting participants based on specific characteristics or criteria. 👉Snowball Sampling: Selecting participants through referrals from initial participants. The choice of sampling method depends on the specific goals of the AI project, the available resources, and the characteristics of the target population. Choosing the wrong sampling technique can lead to biased results, skewed data, and ultimately, the development of AI models that fail to perform well in real-world scenarios. 💡 Quick Tip: When selecting a sampling technique, consider the diversity and representativeness of the target population, the level of precision required, and the resources available for data collection.  #AI #ArtificialIntelligence #productmanagement #productmanager #DataScience 

8

If you're eager to dive deep into understanding transformers, there's no better way than getting your hands dirty and building one from scratch. In this repository, you'll find a comprehensive breakdown of the transformer model's inner workings, including: 1. Encoder-Decoder Implementation: Dive into the translation project from scratch, unraveling the intricacies of both the encoder and decoder components. 2.Positional Encoding Demystified: Understand positional encoding from scratch, a crucial aspect of transformer architecture. 3.Attention Mechanism Dissected: Explore the attention mechanism from scratch, which forms the backbone of the transformer's power. 4. GPT2 Architecture Unveiled: Delve into building the GPT2 architecture from scratch in PyTorch, gaining insights into its structure and functionality. #transformer #DL #AI #ML #genai #GPT #GPT2 #jobseeker https://lnkd.in/g-MGEyTi

4

1 Comment

الحمد لله ch8 Done : Dimensionality Reduction هل كل ما اقلل الفيتشر هاخد دقة اعلي للموديل؟ طيب هل اختلاف الموديل بنفس عدد Features هيديني نفس ال accuracy? طيب هل الفيتشر ال راحت ينفع ترجع؟ طيب لو رجعت هتكون 100% هي ال اتعملها دروب؟ استخدمت 2 Model : 1-Random Forest Classifier 2 - SGDClassifier  The author suggest that you should train your model in all the trainning set before using Dimensionality Reduction as it may make your system perform slightly worse. It also makes your pipelines a bit more complex and thus harder to maintain just like compressing an image to JPEG can degrade its quality. اتعرفت علي PCA: is a linear dimensionality reduction technique that transforms the data into a new coordinate system such that the greatest variance by any projection of the data comes to lie on the first coordinate (called the first principal component), the second greatest variance on the second coordinate, and so on. واتعرفت علي نوع جديد لل DM وهو Manifold Learning : is a type of non-linear dimensionality reduction that seeks to uncover the low-dimensional structure embedded within high-dimensional data. وشفت نوع من انواعه ال هو Locally Linear Embedding (LLE) ده اختصار للشابتر علي كاجل فيها ملاحظات والكودينج ال في الشابتر

2

#RAG! #RAG 2.0!! Next #SOTA!!! But are they the answer you are looking for? Father of RAG, Contextual AI has come up with RAG 2.0 (Contextual Language Models) which claim to solve the problem of having a disconnected system as in the case of Frozen RAG models. The current RAG systems use three disconnected systems, one for Retrieval (Get relevant data original to your context), one for Augmentation (creating prompts to augment user question with business specific context) and one for Generation (usually an LLM system like ChatGPT). All these distinct systems are trained individually on different datasets. What RAG 2.0 claims is a unification of the three by back propping the weights all the way through the retrieval system. While RAG 2.0 breaks all the records on the benchmarks of RAG, is that the latest and greatest solution you are looking for? For most of the businesses the answer will be NO. There are some fundamental questions to be asked before jumping the gun for a SOTA system: 1️⃣ Frequency of Re-training: How often do you need to update the system to maintain its effectiveness? Frequent re-training can be resource-intensive, impacting your operational workflow and potentially escalating costs. 2️⃣ Costs of Training and Inference: Implementing and running RAG 2.0 involves substantial computational resources, especially during the initial training and subsequent updates. Evaluate the financial implications of these processes and consider whether the benefits justify these expenses. 3️⃣ Adaptability to Changing Data: AI systems rely heavily on the accuracy and relevance of their training data. How well does RAG 2.0 adapt if the foundational data changes? This flexibility is critical for maintaining performance and relevance, especially in dynamic industries. RAG 2.0 might be ideal for scenarios that require cutting-edge AI capabilities, but the overall return on investment may not justify its adoption across all types of businesses. What are your thoughts on integrating such advanced systems into everyday business operations? Have you experienced challenges in implementing high-end AI solutions in your sector? Let's discuss the practical aspects of adopting new technologies in our industries. For more details on RAG 2.0 and its capabilities, check out [Introducing RAG 2.0 by Contextual AI](https://lnkd.in/eUgGcSSD).

9

Best feature of End to End Data Science Live classes - we keep updating the course as per market requirement. This Jan, we added GenAI module. For me, teaching GenAI not equals teaching how to prompt. We started with sequence models and 1. How language can be modelled as a sequence and earliest statistical sequence modelling (LM) on Shakespeare corpus 2. What the naive RNNs succeeded at and what were the drawbacks How LSTMs overcame the long and short term dependencies of text and where they failed. 3. How Seq-Seq models gave raise to encoder- decoder architecture in the context of Machine Translation problem 4. And finally how Seq-Seq models gave cross attention and self attention evolved from there. 5. Finally discussing the details of Transformers architecture along with nuances of encoding text. If you want to understand the LLMs in deeper way, the above step by step guide will help. One of the most gratifying part of building this module is I have to go through many research papers, detailed blogs and accumulated a lot of references for further read. And in a few weeks, we are adding elaborate hands on tutorials to build RAG applications in Python !

18

𝐂𝐡𝐚𝐭𝐆𝐏𝐓 𝟒𝐨 can make 𝐛𝐢𝐠, 𝐮𝐧𝐢𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐭 mistakes 😱! Today, I utilized ChatGPT to calculate the correlation between two binary variables. While it may seem trivial (Fig 1 - blue), the result was unexpected. The tool initially calculated 2.48 (Fig 2 - red), highlighting an error in the square root computation. Upon correction, the accurate correlation coefficient of 0.795 was obtained (Fig 3; green). Here is the link to my GPT chat: https://lnkd.in/djB5Jbra Reflecting on this, the question arises - Can we truly trust ChatGPT for complex tasks when it falters on basic calculations? Perhaps, as I encountered, a cautious approach is beneficial. I would recommend this prompt "𝐂𝐡𝐚𝐭𝐆𝐏𝐓 𝐦𝐚𝐤𝐞𝐬 𝐦𝐢𝐬𝐭𝐚𝐤𝐞𝐬. 𝐂𝐡𝐞𝐜𝐤 𝐞𝐯𝐞𝐫𝐲𝐭𝐡𝐢𝐧𝐠," instead of "𝐂𝐡𝐚𝐭𝐆𝐏𝐓 𝐜𝐚𝐧 𝐦𝐚𝐤𝐞 𝐦𝐢𝐬𝐭𝐚𝐤𝐞𝐬. 𝐂𝐡𝐞𝐜𝐤 𝐢𝐦𝐩𝐨𝐫𝐭𝐚𝐧𝐭 𝐢𝐧𝐟𝐨." Have you encountered similar instances of AI discrepancies? I invite you to share your experiences. Thank you.

63

3 Comments

💪 How about supercharging your RAG by using 32x lesser Memory leading to up to 45x Speedup while maintaining ~96% accuracy? Try Binary Quantization! 👇 🚀 But what is Binary Quantization? Here are some details: - Converts float32 values to 1-bit values, resulting in a 32x reduction in memory and storage usage. - Uses Hamming Distance for efficient retrieval, with just 2 CPU cycles. (super fast) - Rescoring step can boost performance. 🚀 Can't compromise on accuracy but still want a boost up? Then try out Scalar (int8) Quantization: - Maps float32 values to discrete int8 values, representing 256 distinct levels. - Uses a large calibration dataset of embeddings to compute the range and steps. - Can be combined with binary quantization for optimal performance. - Can provide up to 4x memory saving and 4x speedup while maintaining 99% accuracy! 🔗 Article: https://lnkd.in/dCSGuYJy #ai #ml #rag #datascience #quantization #machinelearning #llm #embeddings

23

3 Comments

🌟 Exploring MapReduce and Its Versatile Applications Across various Domains 🌟 In the ever-evolving landscape of big data, MapReduce stands out as a foundational programming model that has revolutionized how we process and manage large datasets. Let's delve into what MapReduce is and explore its diverse applications across various domains! 🔍 What is MapReduce? MapReduce is a programming model designed for processing large data sets with a parallel, distributed algorithm on a cluster. The model simplifies data processing by breaking it down into two main functions: Map: Processes input data and produces key-value pairs. Reduce: Aggregates the key-value pairs produced by the Map function, resulting in a condensed output. 🚀 Applications of MapReduce Across Different Domains: Search Engines Web Indexing: Search engines like Google use MapReduce to index the web, allowing for quick retrieval of search results. Keyword Analysis: MapReduce processes large volumes of search queries to identify trending keywords and improve search algorithms. Social Media- User Activity Analysis: Platforms like Facebook and Twitter utilize MapReduce to analyze user interactions, trending topics, and content engagement. Recommendation Systems: MapReduce helps in building robust recommendation systems by processing user data and suggesting relevant content or connections. Finance- Fraud Detection: Financial institutions employ MapReduce to analyze transaction data for patterns indicative of fraudulent activity. Risk Management: MapReduce models assess and manage risks by analyzing historical data and market trends. Healthcare- Genomic Data Processing: MapReduce aids in the analysis of genomic data, facilitating research in personalized medicine and genetic disorders. Patient Data Analysis: Healthcare providers use MapReduce to analyze patient records, improving treatment plans and outcomes. E-Commerce- Customer Behavior Analysis: E-commerce giants like Amazon use MapReduce to understand customer purchasing patterns and optimize product recommendations. Inventory Management: MapReduce processes sales data to predict inventory requirements and manage stock efficiently. Telecommunications- Network Optimization: Telecom companies analyze call data records with MapReduce to optimize network performance and reduce downtime. Customer Churn Prediction: MapReduce models predict customer churn by analyzing usage patterns and service feedback. 💡 Why MapReduce? Scalability: Efficiently handles petabytes of data by leveraging distributed computing. Fault Tolerance: Ensures data processing continues seamlessly, even if some nodes fail. Simplicity: Abstracts the complexity of parallel programming, making it accessible to developers. Let's embrace the power of MapReduce and unlock new possibilities in data processing! 🌐💡 #MapReduce #BigData #DataProcessing #AI #Technology #Innovation #SearchEngines #SocialMedia #Finance #Healthcare #Ecommerce #Telecommunication

15