# Apr 13, 2024

### Top PhD Programs in ML/NLP
- **Top PhD programs in ML/NLP** demand not just publications but also **strong letters of recommendation (LOR) from well-known professors** and personal connections, reflecting a shift towards a more holistic review of applicants' profiles.

### Quantum Algorithms for Lattice Problems
- **Yilei Chen** introduces a **quantum algorithm** for the **Learning With Errors (LWE)** problem, leveraging **Gaussian functions with complex variances** and **windowed quantum Fourier transform** techniques, aiming to solve LWE in polynomial time.

### Publication Rat Race for PhD in ML
- **PhD programs in ML** at top universities now **require applicants to have multiple first author papers** and strong recommendations from reputable researchers, a standard once reserved for faculty position qualifications in other fields.

### Your LLM Is a Capable Regressor When Given In-Context Examples
- **Large language models (LLMs) like GPT-4 and Claude 3** excel in **linear and non-linear regression tasks**, often **surpassing traditional supervised methods** such as Random Forest and Gradient Boosting, especially noted in the challenging Friedman #2 dataset.

### Multiple First-Author Papers in Top ML Conferences, but Still Struggling to Get into a PhD Program. What Am I Missing?
- Despite **multiple first-author papers** in top ML conferences and significant contributions to open-source projects, the author faces **rejection from PhD programs**, questioning the sufficiency of a strong research background without connections.

### NeurIPS 2024 Adds a New Paper Track for High School Students
- **NeurIPS 2024 introduces a new paper track** specifically for high school students, focusing on **machine learning for social impact**, a move that underscores the conference's commitment to nurturing young talent and broadening the scope of contributions to the field.

### Pylyzer – A Fast Static Code Analyzer and Language Server for Python
- **`pylyzer`** is a **static code analyzer and language server for Python**, boasting **superior performance** and **detailed analysis capabilities**, notably outperforming pytype and pyright in speed by over **100 times**.

### Show HN: Stream of Consciousness – Watch an Artificial Persona Making Art
- The **Stream of Consciousness** project leverages **realtime visualization and data streaming** to create an interactive experience.

### Self-Hosted Is Awesome
- **Self-hosting offers significant cost savings and data control**, with modern tools boasting user-friendly designs and robust functionality, contrasting with the rudimentary offerings of a decade ago.

### TwoMinutePapers - DeepMind’s New AI Saw 15,000,000,000 Chess Boards!
- DeepMind's new AI, **without using self-play or search strategies**, learned chess from Stockfish by analyzing **15 billion board states**, achieving grandmaster-level play.

### I Accidentally Built a Meme Search Engine
- Harper Reed developed a **meme search engine** using **siglip/CLIP and vector encoding** to transform images into searchable vectors, inspired by a simple app that utilized similar technology for image similarity searches. [siglip paper](https://arxiv.org/abs/2303.15343)

### When an Antibiotic Fails: Scientists Are Using AI to Target 'Sleeper' Bacteria
- **Scientists at MIT** are leveraging **AI to discover antibiotics** that can target metabolically dormant bacteria, addressing a significant challenge in treating infections that recur due to bacterial dormancy.

### YannicKilcher - ML News
- **AI 22 Labs introduces Jamba**, a **hybrid SSM Transformer model** that combines the Mamba architecture with attention layers, achieving **long context performance** without increased memory demands.

### MindEye2: Shared-Subject Models Enable fMRI-To-Image With 1 Hour of Data
- **MindEye2**, developed by MedARC in collaboration with Stability AI and various research institutions, **generates high-quality images from brain activity** using a novel approach that requires only **1 hour of fMRI data** for fine-tuning with a new participant's data. [Read the full research paper here](https://arxiv.org/abs/2403.11207).