Shambhavi B Sampath

PhD candidate @BMW Autonomous driving. Hardware/DNN co-design. Environment perception. Efficient Edge deployment.

[Akshay 🚀 @akshay_pachaar]

Top 5 LLM fine-tuning techniques, explained with visuals:

[Atul Kumar @atulkumarzz]

I built a dashboard that knows more about your competitors than their own CEOs.

→ Tracks features
→ Spies on pricing
→ Monitors market shifts
→ Even reads their customer reviews

Took me 27 minutes with AI.
Here’s how you can build yours 👇

[The AI Colony @TheAIColony]

3. Building Effective Agents by Anthropic

Link: https://t.co/y5LmYsizap

[Matt Dancho (Business Science) @mdancho84]

Bayes' Theorem is a fundamental concept in data science.

But it took me 2 years to understand its importance.

In 2 minutes, I'll share my best findings over the last 2 years exploring Bayesian Statistics. Let's go.

[Matthew Berman @MatthewBerman]

11/ Links!

Check out the paper here: https://t.co/QA4hqxV1PF

The open-source code here: https://t.co/SNfXU8o7iL

And my full video breakdown here: https://t.co/uPkGm9scYg

[Sumanth @Sumanth_077]

Hands-On Large Language Models!

This repository contains the complete code examples from the book Hands-On Large Language Models.

It includes notebook examples that cover everything from the introduction to language models to fine-tuning them.

https://t.co/9Ii9NVVLFL

[John Rush @johnrush]

I procrastinate all day long testing new AI tools,

New AI tools, news & updates 🧵 :

[Shambhavi B Sampath @Shambhavi95]

@DateConference The #DATE2025 author notifications are out! https://t.co/dI9eZOUnoH
Such great timing, right before the paper registration deadline for #DAC2025. :)

[Shambhavi B Sampath @Shambhavi95]

@SchmidhuberAI Was literature survey difficult back then with limited access to resources and technology? Is there a possibility that plagiarism could have been unintentional?

[kamran Hassan @Rana_kamran43]

10. Yousician:

Yousician is a music education app that teaches you how to play guitar, piano, bass, or ukulele through interactive lessons and exercises.

[kamran Hassan @Rana_kamran43]

5. Kensho

Provides AI-powered analytics and forecasting tools for market trends and investment opportunities.

[Rasel Hosen @details_with_ai]

8. Checkio

Learn Python and Typescript with this game website.

https://t.co/OeqBv9DzzM

[Philipp Schmid @_philschmid]

5 papers you want to read to understand better how @OpenAI o1 might work. Focusing on Improving LLM reasoning capabilities for complex tasks via training/RLHF, not prompting. 👀

> Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking (https://t.co/ote01ei0hQ) from Stanford

> Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents (https://t.co/ZkGtrzRcV9) from MultiOn/Stanford

> Let's Verify Step by Step (https://t.co/g2iPYVMd6N) from OpenAI

> V-STaR: Training Verifiers for Self-Taught Reasoners (https://t.co/34VLwce0HZ**) from Microsoft, Mila**

> Learn Beyond The Answer: Training Language Models with Reflection for Mathematical Reasoning (https://t.co/BJE5u62B39) from Notre Dam, Tencent

I'm not claiming this is how O1 works, but it helps us better understand it. I'll share summary posts in the coming days. Make sure to follow! 🫡

[Sarah Chieng @MilksandMatcha]

Get instantly notified whenever you, your company, or topics of interest are mentioned online 🤖

> Built with @browserbase and @ValDotTown
> Constantly monitors Twitter, Hacker News, and Reddit
> Sends a slack notification every time your keywords are mentioned

Project created with @alexdphan 🛠️

Full code tutorial below 👇

[Satvik Paramkusham @satvikps]

🤯🤯 You can now create a chatbot on ANY GitHub repo using the Llama 3.1 405B model with @huggingface assistants -- FOR FREE! 💰

This is insane! 🚀

Link: https://t.co/gMpBGxAOKB

@ClementDelangue @julien_c

[Pau Labarta Bajo @paulabartabajo_]

Let's build an AI Coding assistant with Llama3 ↓🧵🦙

[Josep Ferrer @iamjosepferrer]

The Transformer's encoder clearly explained 👇🏻

[Robin Delta @heyrobinai]

Claude 3.5 Sonnet is insane.

And people are cooking up incredible use-cases that weren't possible with GPT-4o.

10 examples you don't want to miss:

[Dr Milan Milanović @milan_milanovic]

𝗙𝗮𝗰𝗲𝗯𝗼𝗼𝗸 𝗰𝗿𝗲𝗮𝘁𝗲𝗱 𝗮 𝘁𝗼𝗼𝗹 𝘄𝗵𝗶𝗰𝗵 𝗱𝗲𝘁𝗲𝗰𝘁 𝗮𝗻𝗱 𝗿𝗲𝗽𝗮𝗶𝗿 𝗯𝘂𝗴𝘀 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗰𝗮𝗹𝗹𝘆

A recent document by Facebook engineers explains how they wrote a tool that can automatically fix bugs. In the paper, they introduced 𝗦𝗔𝗣𝗙𝗜𝗫, an automated tool designed to detect and repair bugs in software. The tool has suggested fixes for six essential Android apps in the Facebook App Family: Facebook, Messenger, Instagram, FBLite, Workplace, and Workchat.

How Does It Work?

𝗦𝘁𝗲𝗽 𝟭: Detect a Crash - Another tool, 𝗦𝗮𝗽𝗶𝗲𝗻𝘇, finds app crashes. When Sapienz identifies a crash, it is logged into a database.

𝗦𝘁𝗲𝗽 𝟮: Identify the Problem—SAPFIX pinpoints the exact line of code causing the issue. It first checks if the crash is reproducible. If it's not, the crash is discarded. It uses a technique called "spectrum-based fault localization" to identify the most likely lines of code responsible for the crash.

𝗦𝘁𝗲𝗽 𝟯: Suggest a Fix - Using predefined templates or code mutations, SAPFIX proposes a solution. After identifying the fault location, SAPFIX attempts to generate a patch. It employs two strategies:

🔹 𝗧𝗲𝗺𝗽𝗹𝗮𝘁𝗲-𝗯𝗮𝘀𝗲𝗱 𝗙𝗶𝘅𝗶𝗻𝗴: SAPFIX uses predefined templates to suggest fixes for common bugs. These templates are designed based on standard developer practices.

🔹 𝗠𝘂𝘁𝗮𝘁𝗶𝗼𝗻-𝗯𝗮𝘀𝗲𝗱 𝗙𝗶𝘅𝗶𝗻𝗴: SAPFIX resorts to a mutation-based system if the template-based approach fails. It systematically applies a series of code mutations to the fault location to generate potential fixes.

𝗦𝘁𝗲𝗽 𝟰: Test the Fix—The proposed solution is tested to ensure its validity. It uses the test cases from 𝗦𝗮𝗽𝗶𝗲𝗻𝘇 to check the patch's validity. If the patch passes all tests, it's considered a good fix. After patch validation, SAPFIX uses 𝗜𝗻𝗳𝗲𝗿 (a static analysis tool) to analyze the proposed fix further. Infer checks if the patch introduces any new potential issues.

𝗦𝘁𝗲𝗽 𝟱: Review - Developers get the final say, reviewing and approving the fix.

Check the entire document in the comments.

Image credits: Facebook.

#technology #softwareengineering #programming #techworldwithmilan #coding

[Shubham Saboo @Saboo_Shubham_]

Claude Sonnet 3.5 can now search the web with just 15 lines of Python code (step-by-step instructions):