Engineer Readings – Telegram
[paper][meta][large concept models]
“The current established technology of LLMs is to process input and generate output at the token level. This is in sharp contrast to humans who operate at multiple levels of abstraction, well beyond single words, to analyze information and to generate creative content. In this paper, we present an attempt at an architecture which operates on an explicit higher-level semantic representation, which we name a “concept”. Concepts are language- and modality-agnostic and represent a higher level idea or action in a flow”

https://scontent-ams4-1.xx.fbcdn.net/v/t39.2365-6/470149925_936340665123313_5359535905316748287_n.pdf?_nc_cat=103&ccb=1-7&_nc_sid=3c67a6&_nc_ohc=_Kelt2jn-pkQ7kNvgFRhMPH&_nc_zt=14&_nc_ht=scontent-ams4-1.xx&_nc_gid=AsqaBwO9TftbIqsIF6KCPA3&oh=00_AYAAj36Wbvgp4TU0V0JPoyHGs-_FesxPYaEwDvdGZcbtNw&oe=676768D2

github: https://github.com/facebookresearch/large_concept_model
[ai][genesis]
Amazing step toward with generating models. You can find multiple videos in the link:
https://genesis-embodied-ai.github.io/
👍1
[ai][llm]
Large scale multimodal agents society

“In this paper, taking e-commerce
scenarios as an example, we present LMAgent, a very large-scale and multimodal agents society based on multimodal LLMs.
In LMAgent, besides freely chatting with friends, the agents can autonomously browse, purchase, and review products, even perform live streaming e-commerce. To simulate this complex system, we introduce a self-consistency prompting mechanism to augment agents’ multimodal capabilities, resulting in significantly improved decision-making performance over the existing multi-agent system. Moreover, we propose a fast memory mechanism combined with the small-world model to enhance system efficiency, which supports more than 10,000 agent simulations in a society. Experiments on agents’ behavior show that these agents achieve comparable performance to humans in behavioral indicators. “

https://arxiv.org/pdf/2412.09237
🌟 Happy New Year, everyone! 🌟

This year has been incredible, and I’m so grateful for each of you—our channel grew to 500+ amazing members! 🎉 Thank you for your support, engagement, and encouragement along the way.

I hope the articles, videos, and books shared here added value to your journey. Let’s make this new year even more inspiring and impactful together. Here’s to learning, growing, and achieving great things in 2024!

Cheers to a fantastic year ahead! 🥂Love you all 💛
15🎉4
[ai][long term memory][paper]

https://arxiv.org/pdf/2501.00663v1
[ai][deepseek][paper]
https://github.com/deepseek-ai/DeepSeek-V3/blob/main/DeepSeek_V3.pdf

Model Summary
Architecture: Innovative Load Balancing Strategy and Training Objective

On top of the efficient architecture of DeepSeek-V2, we pioneer an auxiliary-loss-free strategy for load balancing, which minimizes the performance degradation that arises from encouraging load balancing.

We investigate a Multi-Token Prediction (MTP) objective and prove it beneficial to model performance. It can also be used for speculative decoding for inference acceleration.

Pre-Training: Towards Ultimate Training Efficiency

We design an FP8 mixed precision training framework and, for the first time, validate the feasibility and effectiveness of FP8 training on an extremely large-scale model.

Through co-design of algorithms, frameworks, and hardware, we overcome the communication bottleneck in cross-node MoE training, nearly achieving full computation-communication overlap. This significantly enhances our training efficiency and reduces the training costs, enabling us to further scale up the model size without additional overhead.

At an economical cost of only 2.664M H800 GPU hours, we complete the pre-training of DeepSeek-V3 on 14.8T tokens, producing the currently strongest open-source base model. The subsequent training stages after pre-training require only 0.1M GPU hours.
[lecture notes][andrew ng][machine learning]

https://cs229.stanford.edu/main_notes.pdf