Search: “research”

这可能是今年AI圈最清醒的一条推文， Yann LeCun 是当今 AI 领域最有影响力的科学家之一，深度学习三大教父之一，2018 年图灵奖获得者， 他直接怼了Anthropic CEO Dario的著名言论， Dario说未来一到五年，一半的科技法律咨询金融岗位会被彻底干掉。 LeCun说你错了，而且不仅Dario错了，所有的AI实验室CEO 和一众大佬包括Sam Altman，Yoshua

Yann LeCun argues that the normalized exponential function exp(x)/sum(exp(x)) should be called 'softargmax' rather than 'softmax,' as it represents a soft approximation of the argmax operation rather

Exactly. The number of iterations you do is far more important than the amount of 'thinking' you do. It doesn't matter how intelligent or knowledgeable you are. Reality will always push back. The worl

Human cognitive friction has long been acting as a regularizer for a lot of digital infrastructure. It made software APIs less terrible and codebases less complex. Now LLM disintermediation is causing

I'd say cognitive friction is even more than just a regularizer: it's an incentive to find the right interface abstractions, and in turn good abstractions are what enables compounding over time. Piles

Indeed. Indeed. The history of tech impact on labor is well-documented, including by those named. It's unpredictable, but usually improves productivity and leads to expansion. Law & white-collar worke

Just got access to Grok 4.3 and the first thing I did was throw the hardest task I usually run on claude to test it gave it a complex research prompt on deepseek: >thought for 13 mins and 17 seconds >

Automated lint: 43 errors, 9 warnings, 3 info

This post draws a humorous parallel between using images of salarymen eating ramen as a test case for AI image generation models in Eastern contexts and the Western 'Will Smith eating spaghetti' meme,

The two most repetitive unfounded concerns about technology in history are: - Will it cause mass unemployment? - Will it ruin a generation? You don’t get to just say “it’s different this time” and mov

Unlimited demand for a product or service does not guarantee business viability or economic sustainability, as other critical factors such as cost structure, scalability, profitability, and resource c

Video of my opening keynote for the 2026 World Modeling Workshop at Mila, Quebec AI Institute: simple but powerful ways of using world models and their latent space (sorry for sound problem 0:20-0:43)

Constraints are the catalyst of invention. An infinite search space leads to paralysis. The most creative inventions happen when you are forced to solve a problem within appropriately narrow constrain

if you are the only one who can build this amazing tech and believe it is super dangerous, it’s very easy: you quit for the sake of humanity. if you are not the only one who can do it, you are definit

Indeed. The history of tech impact on labor is well-documented, including by those named. It's unpredictable, but usually improves productivity and leads to expansion. Law & white-collar workers aren'

This article outlines Sebastian Raschka's systematic approach to learning and understanding Large Language Model (LLM) architectures, likely covering foundational concepts, key components, and methodo

Mooncake is the serving platform for Kimi developed by Moonshot AI, featuring a KVCache-centric disaggregated architecture that separates prefill and decoding clusters while leveraging underutilized C

The tensor engine was first implemented inside SN3 (before it was called Lush) in 1992 at Bell Labs by Léon Bottom and me. The naming convention has survived to this day in PyTorch and other libraries

Francois Chollet compares programming language philosophies between PHP and Go, likely discussing design trade-offs, developer experience, or use cases that differentiate the two languages. The compar

Francois Chollet discusses a hiring signal for identifying high-quality deep learning candidates based on their choice of machine learning frameworks, suggesting PyTorch proficiency (or lack thereof)

You cannot think your way to a perfect design. Only building and testing, over many iterations, can reveal the flaws in your mental model and provide the feedback you need to create the best design po

arXiv:2601.08831v5 Announce Type: replace Abstract: Video object segmentation methods like SAM2 achieve strong performance through memory-based architectures but struggle under large viewpoint changes

arXiv:2601.04588v2 Announce Type: replace Abstract: Segmentation of the left atrial (LA) wall and endocardium from late gadolinium-enhanced (LGE) MRI is essential for quantifying atrial fibrosis in pa

arXiv:2604.14868v1 Announce Type: new Abstract: 4D millimeter-wave (mmWave) radars are increasingly used in robotics, as they offer robustness against adverse environmental conditions. Besides the usu

arXiv:2604.14263v1 Announce Type: cross Abstract: Accurate detection and segmentation of glomeruli in kidney tissue are essential for diagnostic applications. Traditional deep learning methods primari

arXiv:2604.15215v1 Announce Type: new Abstract: We present a novel hierarchical spatiotemporal action tokenizer for in-context imitation learning. We first propose a hierarchical approach, which consi

arXiv:2510.13829v3 Announce Type: replace Abstract: As large language models (LLMs) continue to advance rapidly, reliable governance tools have become critical. Publicly verifiable watermarking is par

arXiv:2601.16870v2 Announce Type: replace Abstract: Integrated control of wheelchairs and wheelchair-mounted robotic arms (WMRAs) has strong potential to increase independence for users with severe mo

arXiv:2604.14484v1 Announce Type: new Abstract: Behavior cloning (BC) policies on position-controlled robots inherit the closed-loop response of the underlying PD controller, yet the effect of control

arXiv:2509.06593v2 Announce Type: replace Abstract: Accurate odometry is a critical component in a robotic navigation stack, and subsequent modules such as planning and control often rely on an estima

arXiv:2604.14603v1 Announce Type: cross Abstract: The fundamental limit of natural signal compression has traditionally been characterized by classical rate-distortion (RD) theory through the tradeoff

arXiv:2509.24886v3 Announce Type: replace Abstract: Canonicalization is a widely used strategy in equivariant machine learning, enforcing symmetry in neural networks by mapping each input to a standar

arXiv:2601.07667v2 Announce Type: replace Abstract: Due to the prevalence of large language models (LLMs), key-value (KV) cache reduction for LLM inference has received remarkable attention. Among num

arXiv:2604.14779v1 Announce Type: cross Abstract: In continual visual question answering (VQA), existing Continual Learning (CL) methods are mostly built for symmetric, unimodal architectures. However

arXiv:2604.13812v1 Announce Type: new Abstract: Quantum circuit optimization is a central task in Quantum Computing, as current Noisy Intermediate Scale Quantum devices suffer from error propagation t

arXiv:2604.15114v1 Announce Type: cross Abstract: We propose a novel amortized optimization method for predicting optimal transport (OT) plans across multiple pairs of measures by leveraging Kantorovi

arXiv:2604.15171v1 Announce Type: new Abstract: Recent work has shown that diffusion models trained with the denoising score matching (DSM) objective often violate the Fokker--Planck (FP) equation tha

arXiv:2604.14233v1 Announce Type: cross Abstract: The IEC-61850 GOOSE protocol underpins time-critical communication in modern digital substations but lacks native security mechanisms, leaving it vuln

arXiv:2601.06559v2 Announce Type: replace Abstract: Grounding events in videos serves as a fundamental capability in video analysis. While Vision Language Models (VLMs) are increasingly employed for t

arXiv:2604.14755v1 Announce Type: new Abstract: Early identification and removal of polyps can reduce the risk of developing colorectal cancer. However, the diverse morphologies, complex backgrounds a

arXiv:2604.14739v1 Announce Type: new Abstract: Large-scale renewable energy deployment introduces pronounced volatility into the electricity system, turning grid operation into a complex stochastic o

arXiv:2604.14450v1 Announce Type: new Abstract: Quick and accurate emergency handling in Disaster Decision Support Systems (DDSS) is often hampered by network latency and suboptimal application accura

arXiv:2604.14246v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) models have achieved remarkable scalability, yet they remain vulnerable to hallucinations, particularly when processing

arXiv:2601.21003v2 Announce Type: replace Abstract: Large Language Models usually put more emphasis on accuracy and therefore, will guess even when not certain about the prediction, which is especiall

arXiv:2604.14914v1 Announce Type: new Abstract: Text-driven inversion of generative models is a core paradigm for manipulating 2D or 3D content, unlocking numerous applications such as text-based edit

arXiv:2604.15069v1 Announce Type: new Abstract: Graph Neural Networks (GNNs) conventionally rely on standard Laplacian or adjacency matrices for structural message passing. In this work, we substitute