MEMGPT
Create perpetual chatbots with self-editing memory.
2023.10.26
Search the archive
Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Results
16,636 matchesTHE HUMANE AI PIN IS POWERED BY GPT-4
2023.10.26
GENERATIVE AI
2023.10.25
OPENAI PLANS MAJOR UPDATES TO LURE DEVELOPERS WITH LOWER COSTS
2023.10.25
COMPARABLE TO CLIMATE CRISIS
2023.10.25
IMPROVED VIDEO DIFFUSION WITH NOISE SCHEDULING
Generating a single image from a single prompt is generally fine, but fails when generating video because of the temporal changes between frames. However, consistency when changing text is quite challenging. This work tackles both of these problems and enables generations up to 512 frames long using diffusion.
2023.10.25
INTELLIGENT AI ROBOT TRAINING
2023.10.25
USING SEGMENT ANYTHING MODEL FOR 3D MEDICAL IMAGING
SAM-Med3D is an upgraded version of the Segment Anything Model (SAM) tailored specifically for 3D medical imaging. While the original SAM struggled with 3D medical images, SAM-Med3D, trained on a vast dataset of over 131K 3D masks, delivers superior results by capturing 3D spatial details with fewer input prompts.
2023.10.25
ENHANCED VISUAL REASONING WITH VISION-LANGUAGE MODELS
The repo introduces Cola, a system that uses a large language model to coordinate various vision-language models (VLMs) for improved visual reasoning.
2023.10.25
AGENTTUNING
Instruction-tune LLMs using interaction trajectories across multiple agent tasks.
2023.10.25
DEEPSPARSE
DeepSparse is a CPU inference runtime that takes advantage of sparsity to accelerate neural network inference.
2023.10.25
CAN OPENAI WIN CONSUMER AND ENTERPRISE?
2023.10.25
EMOTIONAL LABOR IS HOW WE DIFFERENTIATE FROM AI
2023.10.25
PROTOTYPING FEATURE, STUBBS
2023.10.25
APPLE COULD SPEND $4.75B ON AI SERVERS IN 2024
2023.10.25
CLAUDE AVAILABLE IN 95 COUNTRIES
2023.10.25
AS SOON AS IOS 18
2023.10.24
INTRODUCING THE FOUNDATION MODEL TRANSPARENCY INDEX
2023.10.24
2023.10.24
EVALUATING CHATTING ABILITIES OF LLMS IN MULTI-TURN DIALOGUES
This study assesses the ability of large language models (LLMs) to engage in human-like multi-turn conversations.
2023.10.24
SALMONN: THE AI MODEL TUNED TO UNDERSTAND THE WORLD OF SOUND
This paper introduces SALMONN, a unique AI model that combines text understanding with the ability to process various audio inputs, from speech to music. Not only does SALMONN perform well in trained audio tasks like speech recognition, it also shows impressive emergent skills like translating speech into untrained languages.
2023.10.24
MAKE A FINE-TUNE.JSONL PROGRAMMATICALLY
2023.10.24
A NEW APPROACH TO LIDAR PANOPTIC SEGMENTATION
This project presents a novel detection-centric network for lidar panoptic segmentation (LPS) and tracking, moving away from the traditional segmentation-focused methods.
2023.10.24
IMPROVING OBJECT LOCALIZATION WITH A NEW METHOD
Weakly supervised object localization helps find objects using only simple picture labels. The new Background Activation Suppression (BAS) method does this better by focusing on important parts of a picture and ignoring the rest, showing amazing results on several datasets.
2023.10.24
HOLLY ELMORE AND ROB MILES DIALOGUE ON AI SAFETY ADVOCACY
2023.10.24
EVERY APP THAT ADDS AI LOOKS LIKE THIS
2023.10.24
YOU DO?
2023.10.24
OPENAGENTS, PLUGINS BUT OPEN
Framework for writing language model plugins that gives you full control.
2023.10.24
TO BLOCK GOOGLE
2023.10.24
PI WEB BROWSING
2023.10.24
NEW NVIDIA AI AGENT CAN TRAIN ROBOTS
2023.10.23
OPENAI IN TALKS FOR DEAL THAT WOULD VALUE COMPANY AT $80 BILLION
OpenAI is reportedly nearing a deal led by Thrive Capital that would push its valuation to $80 billion, tripling its value in under six months. This surge in its valuation started with OpenAI's ChatGPT launch, igniting significant investments in leading AI companies. AI entities remain investor favorites despite waning start-up funding due to their transformative potential in tech.
2023.10.23
UNIVERSAL MUSIC SUES ANTHROPIC OVER AI-GENERATED LYRICS
Universal Music is suing AI start-up Anthropic for copyright infringement, alleging the chatbot "Claude" reproduces copyrighted lyrics without permission. The music industry is contending with AI's capability to create "deepfake" songs.
2023.10.23
‘MIND-BLOWING’ IBM CHIP SPEEDS UP AI
2023.10.23
DEEP DIVE: BUILDING LLM-POWERED WEB APPS
2023.10.23
MAKING AI FORGET SAFELY WITH A NEW METHOD
This project introduces a new method called 'saliency unlearning' (Salun) that improves how AI systems forget data.
2023.10.23
CAPTION UP SAMPLING
One of the main drivers of quality for the new DALLE-3 image generation model is improved data quality. This was obtained via caption upsampling with internal models. Here is an implementation with strong open-source models.
2023.10.23
STATE OF MARK PROMPTING
With language models becoming multimodal, we now have more ways than ever to prompt them. State of Mark prompting runs images through Segment Anything first and numbers the detected classes. It then uses that to prompt visual models.
2023.10.23
SOLIDGPT
Chat with and learn from your code repository.
2023.10.23
MATH IS HARD
2023.10.23
NVIDIA BRINGS GENERATIVE AI COMPATIBILITY TO ROBOTICS PLATFORMS
2023.10.23
DALLE-3 RESEARCH PAPER
This paper outlines the main ingredients to DALLE-3's amazing performance. They include a synthetic caption generator, improved modeling with latent diffusion, and improved metrics for faithfulness, style, and coherence.
2023.10.20
FUYU-8B: A MULTIMODAL ARCHITECTURE FOR AI AGENTS
Fuyu-8B, a multimodal model designed for digital agents, is now available on HuggingFace. Unlike other multimodal models, it has a simplified architecture and supports arbitrary image resolutions, responding to large images in under 100ms. Though tailored for specific applications, Fuyu-8B still excels at standard image understanding benchmarks.
2023.10.20
$13M SEED FOR MULTIMODAL SEARCH
2023.10.20
THE WORLD OF NLP: A COMPREHENSIVE TAXONOMY
2023.10.20
TACKLING REAL-WORLD FACE ANTI-SPOOFING CHALLENGES
Researchers present DGUA-FAS, a method designed to address anti-spoofing challenges, especially when faced with changes in domain and unfamiliar attack types. It uses a Transformer-based feature extractor and a special network to simulate unknown attacks. This approach proves more effective in handling both known and unknown threats.
2023.10.20
BENCHMARKING LLM’S GRASP OF FACTUALITY
Researchers introduce FACTCHD, a benchmark designed to detect when large language models, like ChatGPT, produce factually incorrect information, especially in complex scenarios.
2023.10.20
WAYMO SIMULATOR
Waymo has introduced its Waymax simulator for evaluating the performance of agents on self-driving tasks. It is written entirely in Jax.
2023.10.20
XAGENT
XAgent is an open-source experimental LLM-driven autonomous agent that can automatically solve various tasks.
2023.10.20
MOJO🔥 AVAILABLE ON APPLE SILICON
2023.10.20