Category
Transformer、GNN、軽量化など、モデル実装や既存モデル改良に関係する深層学習技術を追います。
🤗 Transformersは、テキスト・ビジョン・音声など複雑なモデル定義をサポートするフレームワークで、インフェレンスターやトレーニングに使用できる。
この研究では、COVID-19臨床パスウェイズの予測監視を支援するために、パイプラインを構築しました。このパイプラインには、データリフティング、時間的再構成、イベントログの構築、プリフィックスベースの表現、予測モデルの整
このリポジトリでは、私的なAIプラットフォームであるDocGPTを提供しています。
🤗 Transformersは、テキスト・ビジョン・音声など複雑なモデル定義をサポートするフレームワークで、インフェレンスターやトレーニングに使用できる。
AIの入門コースを提供する。
分散トレーニングと推論を容易、効率的に実行するためのディープラーニング最適化ライブラリです。
🤗 Transformersは、テキスト・ビジョン・音声など複雑なモデル定義をサポートするフレームワークで、インフェレンスターやトレーニングに使用できる。
AIの入門コースを提供する。
分散トレーニングと推論を容易、効率的に実行するためのディープラーニング最適化ライブラリです。
マシンラーニングシステムの理論と実装に関する本。
オープンソースのAI推論最適化と展開用ツールキットです。
SGLangは、大規模言語モデルのサービングフレームワークです。このライブラリは、高性能なサービスフレームワークで、大規模言語モデルのサービングをサポートしています。
zenmlは、データパイプラインからエージェントまで、AIプラットフォームです。
この論文では、ディフュージョンモデルの高速化を目的としたNVIDIA FastGenについて説明しています。FastGenは、ディフュージョンモデルから高速に生成することが可能です。
このリポジトリでは、私的なAIプラットフォームであるDocGPTを提供しています。
transformerlab-appは、AI研究者に、ローカルハードウェアとGPUクラスタ間でモデルを平滑にトレーニング、評価、およびスケールさせることができる、オープンソースの研究環境です。
Neural Network Inference用の高性能フレームワークです。モバイルプラットフォームに最適化されています。
トークナイザーの高速化を目的としているライブラリ。
Digital Twins are evolving from passive virtual representations into closed-loop systems that continuously syn
Standard spatiotemporal graph convolutional networks need a fixed-size adjacency matrix, coupling model weight
Low-rank adaptation (LoRA) has become a widely used parameter-efficient fine-tuning method for large pretraine
Under expert parallelism, mixture-of-experts (MoE) layers pay all-to-all token dispatch, grouped GEMMs, and a
Predicting stock returns requires modeling persistent company relations and time-varying market signals. Graph
Set separation by continuous linear functionals is fundamental in functional analysis and convex optimization
Wearable sports sensors can recognize actions on the device itself, where lower-precision numbers reduce model
Temporal-difference (TD) learning trains neural representations using bootstrapped targets generated by evolvi
Diffusion models represent a powerful class of generative models known for their solid theoretical foundations
Looped transformers apply a shared block several times per token to improve accuracy without adding parameters
Under temporal nonstationarity, model performance rankings reverse across training horizons due to heterogeneo
Vision-language models (VLMs) increasingly support construction hazard monitoring, yet their frequent false al
What makes a pretraining corpus transferable? We describe a corpus by three properties: its structure, its dim
Individual-level trajectories are essential for applications that rely on user-level mobility behavior. Existi
Discrete diffusion models with parallel decoding are often distilled using token-wise KL objectives. Although
LLM-powered semantic queries increasingly combine relational processing with multiple LLM-based operators over
Power-grid customer service increasingly relies on community social media, where customers describe outages, b
We present an exact computational proof that the frozen QPLIB_2096 instance has global minimum 7,025,720. The
Recent multimodal agents have shown promising progress in embodied planning and web interaction. However, exis
When an AI agent executes an underspecified task, it embeds choices that may contradict unarticulated user con
Mainstream sequence-processing AI is temporally reactive: each update is driven by input and learned state tra
Accurate prediction of Prostate Imaging Reporting and Data System (PI-RADS) scores from multiparametric MRI (m
AI-generated images have reached unprecedented realism, while instruction-driven AI editors re-synthesize imag
Large language model (LLM) services face rapidly rising serving costs. To sustain free or affordable access fo
Large language models (LLMs) often generate responses longer than necessary, increasing inference latency and
Long-range 3D object detection remains difficult for autonomous vehicles because LiDAR returns become sparse a
High-resolution range profiles (HRRPs) provide compact radar target representations but exhibit strong aspect
Foot ulcers are a leading cause of lower limb amputation, and the highest burden often falls on low and middle
Streaming 3D reconstruction must maintain geometric quality over long inputs within a fixed GPU budget. A pret
Self-attention has become the standard for modern sequence modeling, enabling Transformers to model long-range
Large-scale astronomical surveys define a heterogeneous, partially observed multi-survey learning setting: eac
Large 3D repositories make it easy to collect assets, but difficult to determine whether they can be reused re
このリポジトリでは、AIエンジニアリングのためのオープンソースプラットフォームであるMLflowを提供しています。
このリポジトリでは、64MパラメータのGPTを完全にTrainingし、2時間以内に完成させる手法を提供します。
OpenRLHFは、Ray上に構築された強化学習フレームワークです。このフレームワークは、PPO、DAPO、REINFORCE++など、様々な強化学習アルゴリズムをサポートしています。
SANAは、高解像度画像生成モデルSANAを紹介する本研究であり、低計算コストで優れた高解像度画像を生成できる。
FastVideoは、加速されたビデオ生成用の統合推論とポストトレーニングのフレームワークです。
デベロッパー向けのモデロプティミゼーションフレームワークです。モデルの高速化と効率化を実現することができます。
PyTorchで使用できる画像エンコーダとバックボーンの最大のコレクションです。トレーニング、評価、推論など様々なスクリプトや事前の重み付きデータが含まれます。
Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a singl
パラメータ効率の向上のための最先端のフィネチュニングフレームワークです。
電気生理信号から表現を学習し、脳コンピューターインターフェースの開発を支援する。
このリポジトリでは、トークナイザーの最適化を提供しています。
LLMやVLMのFine-Tuningを簡素化したライブラリ。
LLMを利用するために、セマンティック検索やLLMのオーケストレーションなどを行えるフレームワーク。
Selecting small, effective instruction-tuning datasets requires identifying which aspects of the original data
Knowledge distillation is a technique used to transfer the expertise of large teacher models into small, effic
Conformal prediction (CP) provides marginal coverage guarantees, but marginal coverage does not determine wher
Transolver has recently gained widespread adoption for solving diverse physical problems on unstructured geome
In speculative decoding, parallel drafters such as DFlash make drafting cheap, so target verification dominate
We study score learning from denoising pairs when the data contain a hidden non-Gaussian component. If covaria
Speculative sampling can accelerate decoding without changing the target distribution. Its standard guarantee
Offline multi-task reinforcement learning (MTRL) aims to learn a unified policy from heterogeneous offline dat
To enable effective generalization in offline meta-RL, it is essential to accurately infer the underlying task
Long-context inference is bound by the KV cache, which 2-bit codes could shrink about sixfold if their error s
Large language models (LLMs) read company news at scale, and backtests suggest that their readings predict sto
This paper develops a mean-field formation transport method with low online computational cost for large-scale
Probing frozen representations layer by layer is now standard practice, but reporting when a signal first beco
Evolution equations with time-dependent coefficients force a neural solver to spend capacity on a scale that t
Many causal inference problems require sampling from interventional outcome distributions rather than estimati
Large language models (LLMs) can memorize and reproduce fragments of their training data, raising privacy and
Multi-teacher on-policy distillation consolidates expert capabilities by supervising student-generated traject
A substantial line of work builds rotation invariance directly into point cloud architectures. We benchmark 10
Generative AI tools for research are widely deployed, yet individual productivity gains do not guarantee bette
Fast Dominance Estimation Clustering (FastDEC) is an efficient variant of Density Peak Clustering (DPC) that a
Quantum reservoir computing (QRC) processes temporal information through quantum dynamics, but it remains uncl
Large foundation models often answer even when the available visual or textual evidence is insufficient to sup
Multimodal language models continue to improve on broad music understanding benchmarks, but which musical feat
GUI web agents require realistic and interactive environments that remain stable during both evaluation and tr
Muon has recently emerged as a promising optimizer for pretraining large language models, offering a speedup o
Vision-Language Models (VLMs) hold promise for zero-shot video task-progress estimation in robot learning, yet
Evaluating a generative model on all queries from all available benchmark tasks is prohibitively expensive. As
Increased computing power has transformed computational science and improved predictive capabilities for compl
Neural operators provide efficient surrogates for computational fluid dynamics (CFD), but most are designed fo
Most malware detection research focuses on identifying whether a file is malicious before it executes. Far les
Most parameters of a Mixture-of-Experts (MoE) model belong to its experts, and standard training keeps every e
Exemplar-free continual learning on pre-trained Vision Transformers increasingly relies on Low-Rank Adaptation
Computer programmers chain together the same reusable computational primitives, for example, conditional state
LLM agents for GPU kernel optimization sit at two ends of a structure-freedom tradeoff. Scripted agents fix wh
Parameter-efficient continual learning on pre-trained Vision Transformers largely relies on a costly premise:
Graph forecasters for sensor networks are trained in one regime and deployed in others, where sensor coupling
The shear piezoelectric coefficient $e_{15}$ governs the shear-mode electromechanical coupling in piezoelectri
Convexity has three equivalent descriptions: support by affine functions, monotonicity of the subdifferential,
Machine unlearning seeks to remove the influence of deleted training data without full retraining. In the sequ
We study the gradient query complexity of sampling from a density $π(x)\propto\mathrm{e}^{-V(x)}$ on $\mathbb{
Generative models increasingly simulate users at different skill levels in education and AI evaluation. Yet it
Despite their widespread success across machine learning, the learning dynamics and generalization properties
Optimal control of partially observable Markov decision processes (POMDPs), in general, requires a complete hi
Spherical-harmonic neural operators provide an efficient representation for global forecasting, but their disc
Super-resolution networks are trained for one magnification factor, so several output resolutions require seve
Identifying ligands for a protein target requires searching a vast chemical space for a comparatively small se
Robust machine learning for graph-structured data has made significant progress against test-time attacks, yet
Attention remains a substantial computational cost in Transformer training, particularly at long context lengt
Generative AI is billed per token, yet tokens are not equally expensive. Each token enters the KV-cache and ra
A model pretrained to predict a customer's next event can estimate the probability of future events without tr
As general-purpose Large Language Models (LLMs), such as Gemini and DeepSeek, exhibit increasingly advanced re
Zeroth-order (ZO) optimization enables memory-efficient large language model (LLM) fine-tuning, and recent met
Large-scale annotation of environmental sound recordings presents significant challenges under time and resour
Zeroth-order (ZO) optimization has emerged as a promising approach for memory-efficient large language model (
Anomaly detectors rank unfamiliar examples for expert confirmation, and the confirmed examples teach a classif
Dynamic graphs capture evolving interactions in Web systems, including online communication, user engagement,
Responses with headers, lists and greater length receive high scores from reward models (RMs), and RLHF polici
Laminar-turbulent transition governs skin friction, heat transfer and losses in turbomachinery, yet Reynolds-a
Understanding how stochastic gradient noise affects the convergence of Adam is an important problem in stochas
Spherical transformer configurations can accelerate grokking through simultaneous modifications of several mod
Battery degradation remains a central challenge in the development and broad application of sustainable energy
Large language models (LLMs) have introduced a generative approach to recommendation, producing item lists fro
We classify real and complex state-space models (SSMs) with linear transitions, distinguishing fixed-basis dia
Claim verification is the task of deciding whether a source text supports a claim, contradicts it, or lacks su
Beamforming designs for integrated sensing and communication (ISAC) in cell-free multiple-input multiple-outpu
Adaptive compressed sensing (CS) aims to select the most informative measurements sequentially, which is cruci
Multimodal fusion is studied under hybrid federation, where data are partitioned across samples and feature bl
Most fuel oil consumption (FOC) models for voyage planning are per-ship models and thus generalize poorly beyo
Language models are typically trained with a single choice of architecture on diverse domains and data sources
Generative recommendation has been widely adopted in large-scale industrial recommendation systems, where mult
Block-scaled lookup-table (LUT) formats, which represent weights as indices into a small value set (the grid),
Representation alignment (REPA) is widely used to accelerate the training of latent-space diffusion transforme
Given a relational database (RDB) storing heterogeneous tabular information, how can we predict missing (or fu
Representation regularization has emerged as an effective way to improve diffusion transformer training. Yet o
In the prediction and control of physical systems, combining records from different experiments and operations
Many control tasks require driving the system into a prescribed goal set and subsequently keeping it there. Th
The Poll Schumann model requires airframe and engine parameters, but published tables cover only part of the f
We establish mathematical foundations for optimal spatiotemporal vector fields governing continuous score-guid
Bayesian online changepoint detection provides a natural framework for quantifying uncertainty about whether a
We study the spectral structure of the self-attention matrix in Transformer models through an idealized spectr
We give a proper membership-query algorithm for agnostically learning halfspaces under the uniform distributio
As large language models are fine-tuned for increasingly specialized tasks, their performance depends more on
Serving an LLM efficiently depends on the engine, how the model is split across GPUs, precision, scheduler set
Causal inference has emerged as a powerful paradigm enabling robots to make safe, efficient decisions by reaso
Modern learning systems increasingly combine pretrained models with downstream heads. Despite the growing impo
Translating population-level exercise guidance into individualized prescriptions requires integrating clinical
Healthcare information systems are fragmented, and retrieving data from legacy web portals is a recurrent need
Reasoning substantially improves language models' ability to solve difficult problems, but long reasoning trac
Long-to-Short (L2S) model merging seeks to combine the strong reasoning capability of long chain-of-thought mo
Next-visit diagnosis prediction forecasts the diagnoses documented at a patient's next encounter, turning retr
Timely and accurate crop health diagnosis is essential for sustainable agricultural production, yet accurate d
Skin cancer diagnosis, particularly differentiating melanoma from benign nevi, is challenging due to the visua
Skill extraction benchmarks score models against a few human-annotated positives and treat every other skill i
Matrix multiplication dominates both PHY signal processing and Deep Neural Network (DNN) inference, yet these
Language models in customer-facing products must be continually updated, since the data they serve shifts over
AI benchmarks have made remarkable progress in measuring the ability of AI systems to perform increasingly dem
Respiratory sound classification from auscultation recordings is confounded by heterogeneous acquisition condi
Instruction tuning for multi-task network traffic analysis requires data with clearly defined sources, supervi
We introduce Axion 1.0, a fully open agent foundation model that combines strong and broad agentic capabilitie
Rapid advances in cowork agents are shifting human-AI collaboration toward real-world automation, delivering s
Repository-level AI coding benchmarks measure task success but do not directly capture the conditions required
Despite recent advances in CUAs, their development remains constrained by the limited diversity, controllabili
Large language models (LLMs) are driving autonomous agents toward real-world environments, with the Web emergi
Accurate preoperative prediction of lung adenocarcinoma invasiveness is important for surgical planning and in
Theoretical limitations of transformers on state tracking---the ability to track information about entities or
Feedback-driven iteration is widely used to improve LLM-generated scientific ideas, but existing systems eithe
Complex video editing requires understanding complex instructions while jointly fulfilling multiple editing re
Self-supervised learning inherits the coverage of its source data. When an unlabeled corpus is long-tailed, th
Large Language Models (LLMs) have shown promise for empowering search agents to tackle complex information-see
Recent advancements in sign language processing can be attributed to developments in both large-scale dataset
Large Language Models (LLMs) are increasingly used in medical question answering (QA) systems. While recent wo
Speculative decoding (SD) accelerates large language model inference by reducing the sequential cost of token-
Recursive self-improvement (RSI) in production agents involves transforming accumulated experience into reliab
This paper introduces a reinforcement learning-based approximate dynamic programming approach to optimize the
Reinforcement learning (RL) elevates joint audio-video generation, but optimizing heterogeneous, modality-spec
Recursively self-improving (RSI) systems are increasingly viewed as a promising means of accelerating progress
We study the iterates of a fixed ReLU network $N$ of width 4, that is, the discrete-time system $x_{k+1} = N(x
We introduce Dory1.0-4B, a 4-billion-parameter looped (latent-recurrent) language model with elastic depth: a
Activation steering offers a lightweight mechanism for controlling model behavior, but existing methods depend
The widespread adoption of Large Language Models (LLMs) demands reliable detection of human-authored versus sy
People often attach instructions to their search queries, asking for pages written for beginners, documents in
Scaling depth is a key route to stronger large language models, but Pre-LN architectures often suffer from dep
On-policy distillation (OPD) mitigates the student's exposure bias by optimizing on student-generated trajecto
Complex documents such as scientific papers contain rich information in tables, figures, and equations; howeve
We propose a simple, linguistically motivated approach to detecting AI-generated peer reviews based on the inh
The scarcity of human-annotated literary corpora limits the development of domain-specific sentiment analysis
As streaming generation becomes the dominant serving paradigm for both text-only and multimodal large language
Prompt compression is aimed at reducing input prompt lengths to enable cheaper, faster, and potentially higher
Multi-teacher on-policy distillation (MOPD) distills domain experts into one student but needs every expert at
Large language models (LLMs) are increasingly used to evaluate candidates for professional roles, raising conc
When a masked diffusion language model is sampled in a few parallel steps in a random order, the model actuall
Large language models (LLMs) have achieved strong performance across diverse tasks, from question answering to
Steering vectors align a model at inference time without touching its weights, but eliciting several personas
Enterprise conversational systems require memory to personalize interactions, retain information across sessio
Scaled weight decay for Muon (Muon-SW) has been reported to train language models more efficiently than consta
LLM planning often relies on reactive executors (e.g., ReAct) that generate a single action at each step to re
Retrieval-augmented generation (RAG) extends large language models with external evidence, yet relevant eviden
Agents are increasingly used for coding tasks, but the inference cost of large language models (LLMs) limits w
Generating visualisations from natural language requests (text-to-vis) enables users to explore data more effi
Text-to-SQL systems answer natural language questions over structured data, but can fail before SQL generation
Cross-Domain Few-Shot Segmentation (CD-FSS) aims to segment objects from unseen domains using only a few annot
Intracardiac flow is driven by complex, and time-varying cardiac motion and interaction with the heart valves,
Transfer learning from pretrained backbones has become standard practice in multi-sensor computer vision---yet
State-of-the-art bronchoscopy guidance systems and computer-aided diagnosis methods have emerged as important
Longitudinal rodent positron emission tomography (PET) studies of stroke and neuromodulation rely on accurate
Diabetic macular edema (DME) is a leading cause of severe vision loss among working-age adults. Although intra
Feature matching underpins many robotic perception tasks, including visual localization, mapping, SLAM, and au
Automatic rigging is a fundamental task for both static 3D and dynamic 4D mesh animation, yet existing methods
Robust 3D point-cloud recognition is critical for robotics, mapping, and urban sensing, yet point-cloud classi
We introduce CoLa3D, a simple and effective model for composable 3D scene decomposition that, given a single w
High frame rate HDR video is important in applications such as autonomous driving and scientific imaging, wher
Text-to-CAD generation aims to make CAD modeling more accessible by generating executable CAD programs from na
Foundational vision models have become the de facto standard for many vision tasks due to their strong perform
Video dubbing aims to synthesize lip movements that match target speech while preserving the speaker identity,
In this work we present a novel method for the challenging problem of depth image upsampling. Modern depth cam
Streaming digital humans produce detailed video while continuing speech as new text arrives. High-resolution d
Source-only domain generalization (DG) for remote sensing semantic segmentation seeks reliable predictions in
Multimodal retrieval must integrate complementary evidence from views that may be missing, noisy, or unequally
Monocular head pose estimation is usually formulated as absolute regression from a single face crop, yet many
Omni-modal large language models (Omni-LLMs) typically process large numbers of multimodal tokens from audio-v
Class-incremental learning (CIL) for outdoor 3D object detection remains largely underexplored. Existing outdo
Estimating the metric 3D position of people from a monocular camera is important for robot navigation and moti
Multimodal large language models (MLLMs) have shown strong potential in robotic surgery for question answering
Accurate mapping of shallow-water environments using unmanned aerial vehicle (UAV) imagery is challenged by wa
3D Gaussian Splatting (3DGS) has become a vital tool in novel view synthesis. 3D Gaussian functions with attri
Among visual localization strategies, scene coordinate regression (SCR) has emerged as an attractive approach
Multi-source remote sensing image classification benefits from complementary spectral and structural informati
Numerous distillation methods have been developed to accelerate diffusion models. Recent research indicates th
Existing RAW-based detectors typically introduce learnable ISP modules to convert RAW inputs into sRGB-like fo
Three-dimensional human pose estimation has become a fundamental component of computer vision systems for spor
Multispectral imagery provides spectral information beyond the visible range for vegetation monitoring, but it
Classical structure-from-motion reconstruction proceeds iteratively, refining camera poses and scene geometry
Urban photogrammetric point clouds are commonly processed as local 3D patches, limiting the context available
Autoregressive video diffusion models (AR-VDMs) enable real-time, long-horizon video generation but degrade ra
Range-image-based point cloud super-resolution provides an effective way to reconstruct sparse outdoor point c
Few-step autoregressive video diffusion enables efficient generation, yet maintaining consistency over long ro
Lip reading plays a crucial role in applications such as speech recognition, assistive technologies for the he
Face anti-spoofing (FAS) models often learn attack- and dataset-specific cues that degrade performance under u
Learned keypoint descriptors trained on upright imagery often fail under large in-plane rotation. Steerers add
Long-form multi-shot video editing requires an intended edit to remain coherent across abrupt changes in scene
Skeleton-based action recognition is a pivotal task in computer vision; however, conventional Graph Convolutio
Joint nuclei instance segmentation and classification requires precise boundaries and neighborhood information
Jointly trained multimodal networks frequently underutilize weaker modalities, as faster-learning modalities d
Diffusion models are powerful generative frameworks for producing high-quality images by denoising latent vari
Effective polyp segmentation in colonoscopy images is crucial for early detection of colorectal cancer, a lead
We present GazeTags, a camera- and computer-vision-free approach for worn head-gaze object awareness. Our smal
Visual mental imagery (VMI) decoding uses brain activity to reconstruct internally generated visual scenes, pr
3D generation, editing, and completion are typically treated as regenerating the complete 3D asset under task-
Prompt engineering is commonly used to encourage text-to-image models to generate more detailed images. Howeve
Safe autonomous driving requires more than recognizing the current scene: the ego vehicle must anticipate how
Skin cancer classifiers need large sets of annotated images, which are scarce, especially for rare lesion type
Cooperative perception (CP) requires substantial communication bandwidth to exchange sensing information among
Inertial sensors are moving into clothing, from sensorized sportswear to garments instrumented with hundreds o
Hyperspectral image (HSI) classification requires the representation of contiguous spectral signatures and spa
Robot swarms harness simple, local interactions among many individually limited robots to produce complex, coo
Computing distances between sets is important in many robotic applications, such as motion planning and contro
One big challenge in physical AI is interacting with systems whose states are defined by implicit equilibrium.
Mechanical thrombectomy (MT) is an effective treatment for acute ischemic stroke, yet access remains limited b
Myoelectric Control Systems (MCSs) enable voluntary control of wearable robots through muscle activity. Howeve
Contact-rich manipulation requires fine-grained control and complex physical interactions with the environment
Vision-language-action (VLA) models offer a promising path toward general-purpose robot manipulation. While in
Contact-based robotic manipulation is constrained by end-effector reachability and the mechanical properties o
Markerless hand-eye calibration under continuous motion requires jointly resolving camera geometry and the tim
Humanoid loco-manipulation requires coordinating object acquisition, transport, and release from onboard sensi
Robust multimodal perception for soft robotic grasping is challenged by modality-dependent noise, semantic amb
Vision-language-action (VLA) policies can handle noncritical stages of long-horizon manipulation yet fail duri
Long-horizon robotic assembly requires reliable perception and action across grasping, alignment, and insertio
World Action Models (WAMs) have recently shown great promise in robotic control, benefiting from the future im
The spring-loaded inverted pendulum (SLIP) describes the center-of-mass (COM) motion of runners but has no tru
Edge computing service deployment requires the Network Operator (NO) and Service Providers (SPs) to coordinate
This paper proposes a cooperative game-theoretical framework for the financial settlement of real-time deviati
Over-the-Air Federated Learning (OTA-FL) improves communication efficiency by exploiting waveform superpositio
Open Source Computer Vision Libraryは、画像やビデオを分析するためのライブラリです。
🧑🏫 60+ Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers
Faster Whisper transcription with CTranslate2
デバイス上のLLM推論をXビット量化を使用したもの。
Explore concepts like Self-Correct, Self-Refine, Self-Improve, Self-Contradict, Self-Play, and Self-Knowledge,
セマンティックシーケンス分割モデルのライブラリです。
Machine Learning Engineeringは、機械学習の開発と運用をサポートするためのリソースを提供する。
presidioは、テキスト、画像、構造化データを含む敏感データを検出、削除、マスク、アノニマイズするオープンソースフレームワークです。自然言語処理、パターンマッチング、カスタマイズ可能なパイプラインをサポートします。
Accurate forecasting and temporal downscaling of solar irradiance is critical for reliable renewable energy in
When testing a composite null $\mathcal{P}$ against a composite alternative $\mathcal{Q}$, it has been known s
Language and vision models became data-efficient by pretraining on plentiful unlabelled data; physical field m
Data-driven sequential decision making in marketing and healthcare can benefit from observations collected in
We study the implicit bias of Adam and its similarity to sign gradient descent on homogeneous models in binary
We study nonparametric regression using an over-parameterized two-layer ReLU neural network with fixed covaria
Sparse mixtures of experts combine multiple prediction models while evaluating only a small subset for each in
We present HyVox, a hybrid neuromorphic--von Neumann system for low-energy open-set speaker identification und
Machine learning systems are challenging to test using traditional software testing methods because their beha
Standard attention forms query (Q), key (K), and value (V) matrices through learned linear projections of inpu
Bursty serving changes both queue states and the population selected for each model. We study conditional cali
On-policy distillation (OPD) trains a student on its own generations by matching the teacher's token probabili
Effective multi-agent collaboration requires finalizing solutions from reasoning that is distributed across th
Software development sessions often involve several unfinished tasks rather than one issue carried through to
While language models (LMs) excel in natural language tasks, their ability to learn impossible languages--thos
Medical language models promise expert clinical support, but the strongest open-source LLMs remain too large f
On-policy distillation (OPD) can improve answer accuracy while weakening responsiveness to supporting evidence
Linear recurrences are becoming an increasingly common building block in large language models, replacing the
How does a relational representation shared across queries drive a query-specific choice? Controlled entity-bi
Language models can analyze and assess student texts using predefined rubrics and deliver structured feedback,
LLM-based role-playing agents are fluent at imitating a character's style, yet struggle to make role-consisten
Self-evolving agents accumulate experience from task streams, but existing frameworks typically rely on fixed
Web communities rely on moderators to catch harmful messages such as insults, hate speech, and scams, but no o
Electroencephalography (EEG) provides a non-invasive means for neurological and mental health assessment, yet
Depression affects millions of Bengali speakers across South Asia, yet automated detection systems have relied
Verifying that a packaged food is Halal requires reading ingredient declarations and E-number additives before
Representation steering methods like difference-in-means (DIM) offer a lightweight approach to controlling LLM
While online medical consultation broadens access, unprompted patient disclosures and topic shifts frequently
Language agents read code, tool output, and dialogue in contexts that grow at every step. Dynamic sparse atten
Sparse autoencoder (SAE)-based activation steering is a lightweight approach for modifying the behavior of lar
Multimodal Large Language Models (MLLMs) have made substantial progress in 3D scene understanding from multi-v
Vision-Language Models (VLMs) and Vision-Language-Action (VLA) models process hundreds of visual tokens, makin
Task-optimized deep neural networks have risen to prominence as the most predictive phenomenological models of
Repeated satellite imagery permits understanding of how our planet changes over time, letting us predict its f
High-resolution three-dimensional (3D) microstructure characterization remains difficult to undertake at scale
Roadside sensing infrastructure is essential for Intelligent Transportation Systems (ITS), but cameras and leg
Automated classification of retinal pathologies from Optical Coherence Tomography (OCT) images remains challen
The efficiency and fidelity of 3D Gaussian Splatting critically depend on adaptive density control, which grow
Studies of neural network generalization and robustness often summarize smoothness with a single number, such
Robust 4D reconstruction of deformable endoscopic scenes from monocular videos remains challenging due to non-
Reinforcement learning with verifiable rewards (RLVR) has become an effective post-training paradigm for visio
Deep learning models for glaucoma detection are usually tested on images from the same dataset used for traini
This paper introduces a directional bounded-confidence extension of the Cucker--Smale (CS) flocking model that
Dexterous in-hand manipulation is commonly formulated around explicit object pose estimation, yet reliable pos
Mechanical hysteresis and pneumatic loading complicate contact force estimation from deformation measurements
Multi-robot systems in automated warehouses must efficiently attend to a constant stream of pickup-and-deliver
Embodied question answering asks an agent to navigate, observe, and then answer questions about what it saw, y
Most robot policies, whether trained by imitation learning (IL) or reinforcement learning (RL), map observatio
This report presents GatedVLA, our entry for the Physical AI AV (PAI-AV) track. Our learned vision-language-ac
Robotic manipulation requires accurate metric depth, yet commodity RGB-D cameras often fail on non-Lambertian
感覚変換、すなわちバーチャルエキスパートが可能なPytorch実装。
このライブラリは、コンピューター ビジョンのための高度なAI解釈と可視化ソリューションです。このライブラリは、CNN、ビジョン トランスフォーム、分類、物体検出、分割、画像類似度など、さまざまなコンピューター ビジョンの
長時間のビデオ生成を実現するためのモデルのサポートを紹介している。
BERTopicでは、BERTとc-TF-IDFを用いて、トピックを簡単に分析して解釈できるようにしたモデルを紹介しています。
Operators choosing among decision rules, such as when a battery charges or where an idle taxi goes, rank them
In this paper, we establish sharp lower bounds for Gaussian approximation of Polyak-Ruppert averaged stochasti
Interaction trees reuse outcomes to discover treatment-effect heterogeneity, complicating inference after grow
In this paper, we develop a projected instrumental variable framework that relaxes conventional assumptions fo
ALDER, a single-loop augmented-Lagrangian damped extragradient method, is proposed for smooth nonconvex--conca
Distributed support vector machine (SVM) learning is challenging under heterogeneous local data and limited co
Reservoir computing offers a way to study the brain's wiring before it is even trained. The recurrent network
Surprisal correlates closely with reading time and is widely used in the cognitive science of reading as a mea
Diffusion large language models (dLLMs) generate tokens in parallel via iterative denoising, but their bidirec
Character names in large language model (LLM) fiction concentrate on a handful of recurrent names such as 'Ela
Exploring systems that bridge the gap between non living and living matter requires a reconfigurable environme
Diffusion-based vision-language-action (VLA) models cascade pretrained vision-language models (VLMs) with diff
In this paper we present the methodologies, results, and analysis used to deploy a decentralized safety verifi
This paper studies safe policy improvement for real-time operation control of flex-route transit (FRT) systems
This letter introduces HALO-LP, a novel homogeneous all-shifted non-interior-point method for linear programmi
Streaming 3D reconstruction models keep past video frames in a recurrent state, fast weights from test-time tr
Long-horizon embodied manipulation requires policies to act on information that is no longer visible. Modified
Vision-language action (VLA) models excel for end-to-end autonomous driving, yet their tendency to over-invoke
A robot's next action can depend on recent motion or earlier visual evidence absent from its current observati
Laser doppler vibrometer (LDV) scans can be used to detect defects and cracks in manufactured parts before the
Modern robot policies often perform expensive inference to generate action chunks, making serving a latency bo
Deploying deep learning on microcontrollers (MCUs) requires fitting both the model weights and runtime activat
Voluntary platoon formation requires vehicles to coordinate despite uncertain downstream fuel savings and priv
Real-time long-video generation requires continuous output with consistent appearance and natural motion. Exis
Streaming video LLMs can now commentate live gameplay, but they are trained and tested on the same games. We s
データサイエンスの学習には役立つリポジトリ。実世界の問題に応じた学習が可能。
Open data and scalable training for long-horizon video world models.
Performativity arises when prediction-informed decisions change the data-generating process, creating feedback
Phase I dose-finding trials must learn from sparse toxicity observations while limiting exposure to overly tox
In this paper, we consider a class of possibly nonmonotone finite-sum composite inclusions in which evaluating
Estimating the causal effect of a treatment $X$ on an outcome $Y$ from observational data in the presence of u
Low-rank adaptation (LoRA) enables parameter-efficient fine-tuning but can suffer from redundant basis directi
Accurate forecasting with uncertainty quantification is critical in a variety of scientific applications rangi
Reinforcement learning (RL) embodies a foundational paradigm for sequential decision-making in the face of unc
We show that outlier eigenvectors of the Bethe Hessian achieve weak recovery in the stochastic block model dow
Self-distillation methods for large language models with an exponential moving average (EMA) teacher already r
Long-horizon LLM agents play an important role in solving complex and sustained tasks, with ever-growing inter
Gating mechanisms allow a fixed neural network to adapt its computation to the current input by dynamically mo
This paper presents a neural network model for storing and recalling color images of different sizes using the
Tactile sensing provides robots with direct observations of contact physics, local geometry, and interaction d
Humanoid controllers have acquired increasingly diverse motor skills, yet translating these capabilities into
Foundation models are becoming remarkably reliable at high-level reasoning, yet they remain too slow and costl
Autonomous wheelchairs can improve the quality of life for individuals with mobility impairments by providing
Embodied agents are usually organized around individual tasks, yet the physical world keeps changing between t
Simulation benchmarks score robot policies by success rate, and every failed episode counts against the policy
Camera localization methods are mostly multi-stage pipelines relying on dedicated and complex maps that are co
This paper leverages the higher-order kinematic derivatives of serial robots to achieve efficient, aspect-pres
Architectural spaces are increasingly gaining robotic capabilities. Architectural robots such as robotic parti
Recent advances in generative foundation models have expanded the capabilities of embodied AI, enabling increa
Wearable force-augmentation technologies, including exoskeletons and soft exosuits, are being developed for oc
Continual adaptation to new tasks is essential for robots deployed in the real world, but adapting pre-trained
Robot policy learning is dominated by zeroth-order methods such as PPO, which treat the simulator as a black b
Embedding models underpin virtually every modern AI system, from search engines to recommendation systems. Des
Decision models target structured tasks by directly predicting choices or scores, avoiding autoregressive deco
Unlike autoregressive (AR) models, diffusion language models (DLMs) permit parallel token updates, flexible ge
Retrieval-augmented generation (RAG) lets a large language model (LLM) answer questions by retrieving relevant
Linear attention scales linearly with sequence length, but its finite-state memory can limit long-context retr
Dense 3D tracking from monocular video requires establishing point correspondences over time, estimating their
Federated learning (FL) keeps raw data local, but most deployments still rely on a trusted coordinator for cli
この研究では、COVID-19臨床パスウェイズの予測監視を支援するために、パイプラインを構築しました。このパイプラインには、データリフティング、時間的再構成、イベントログの構築、プリフィックスベースの表現、予測モデルの整
AIモデルを高速にトレーニングするためのライブラリ。1台から10000台のGPUで利用可能。
AI用のデータセットを提供するプラットフォームです。
Low-precision training reduces computation but introduces additional rounding error. We theoretically study ho
The marginals of a continuous diffusion model can be reproduced by a deterministic ordinary differential equat
Continual adaptation requires deciding not only how to update a model, but whether to commit an update without
Recent advances in Time Series Foundation Models (TSFMs) increasingly rely on axial attention mechanisms, but
Reinforcement learning (RL) is increasingly used to train robust, adaptable systems across healthcare, e-comme
Probabilistic meshless methods solve partial differential equations (PDEs) by placing a Gaussian process prior
Adaptive inference systems balance accuracy and inference cost by routing inputs to models with varying capabi
We show that the transformer architecture emerges as the score of last-token laws in a natural statistical mod
A central topic in online and stochastic optimization is to design algorithms that do not require tuning of hy
We study adaptive UAV sampling for radio environment map (REM) construction from non-Gaussian IoT observations
Many applications of multi-objective reinforcement learning (MORL) require policies that balance conflicting o
Whether models trained on different modalities converge to the same representation of the world is central to
Rotary Position Embedding (RoPE) typically uses uniform coverage across Transformer layers despite heterogeneo
Pixel-space generative Transformers typically use one token per image patch, so capturing finer detail require
The advent of large, high-density FPGAs such as the AMD Versal enables efficient hardware implementation of ne
Humans and animals have a unique capacity to generalize and learn in novel circumstances, including "0-shot" i
The tension between the established technologies of connectionist AI and their neuromorphic challengers is att
Spiking neural networks (SNNs) promise event-driven, low-power inference, but on vision tasks they still acqui
Formula-Driven Supervised Learning (FDSL) enables vision pre-training without real images, but existing method
This letter proposes MulClip, a lightweight activation function for reservoir computing (RC) on microcontrolle
We study equilibrium computation in multi-unit trading networks with valuations satisfying generalized full su
In a duopoly with captive and contested consumers, I characterize every pair of consumer and producer surplus
Predicting future physical states stands as a central challenge in embodied intelligence. Existing video world
Progress toward artificial general intelligence (AGI) has largely relied on scaling a resource-intensive execu
A capable voice assistant must do more than recognize words; it must understand requests, reason over context,
Recent advances in video world models have enabled the generation of increasingly realistic and interactive vi
Proactive agents must decide whether, when, and how far to assist without an explicit request. Existing evalua
Training acceleration methods for diffusion transformers, such as REPA and dispersive regularization, are comm
Interactive world models generate visual content conditioned on user actions, yet maintaining spatial consiste
The hair of stylized characters in games and avatars is built from hair cards: artist-placed ribbons whose geo
One-step and few-step text-to-image generation must accurately render the details specified in the prompt desp
We present Youtu-Parsing-Omni, an omni-modal parser that converts documents, images, audio, and video into str
Retrieval-augmented generation commonly serves separate models for reranking and answering, duplicating model
While frontier vision-language models demonstrate strong capabilities in user interface (UI) navigation, their
Weight-normalized (WN) training is emerging as a compelling alternative to conventional weight decay (WD), wit
Video diffusion transformers generate high-quality videos, but the quadratic cost of 3D full attention dominat
语言サーバーの開発に使用するライブラリを紹介している。
Pythonで使えるマシンラーニングライブラリを紹介している。
Industrial systems continuously generate time series that vary in sampling frequency, scale, dynamics, and ope
Neural networks are powerful memorizers which are often able to interpolate their training data even when the
Human judgment calls made throughout the data science life cycle can change the conclusions drawn from data. I
Streaming federated learning (SFL) allows clients to train a shared model from continuously arriving data with
We study adversarial linear bandits with strong corruption, in which an adaptive adversary chooses the reward
Kernel methods face two recurrent bottlenecks: the cubic cost of inverting an $n\times n$ kernel matrix, as in
Many problems in computational statistics and machine learning can be formulated as approximating a target dis
This work investigates device-informed spike-timing-dependent plasticity (STDP) in spiking neural networks usi
Energy-constrained aerial platforms require accurate object detection under strict power and latency budgets,
Time-series forecasters are ranked almost entirely by accuracy metrics, but a forecast's decision value, its e
Model distillation has become a major concern in the rapidly growing generative AI discourse. In its simplest
Jev enables efficient decision-making with large language models (LLMs) through typed querying primitives, all
On-Policy Self-Distillation (OPSD), which aligns a student model's token-level distribution with that of its s
Multi-Token Prediction (MTP) is now a common ingredient in language-model pre-training, as it improves downstr
Self-distillation offers a way to improve reasoning models when a suitable external teacher is unavailable. Ho
この研究では、重症看護病棟内の気流と光など環境要因が気性変化の予測に利用できるか検討した。その結果、日照時間と音圧レベルを考慮すると、気性変化の予測精度が向上した。
Deep Gaussian processes (DGPs) are flexible surrogate models for computer experiments with nonstationary respo
Quickest detection problems (QDPs) arise in a wide range of applications. However, the complementarity structu
Graph neural networks (GNNs) rely critically on neighborhood aggregation, yet conventional mean aggregation ma
Language model performance depends strongly on the mixture of domains in the pretraining data corpus. Popular
Accurate forecasting of global total electron content (TEC) is important for space weather monitoring and the
Field applications in chemical sensing, from portable safety monitors to mobile robots, require models that ar
We study how boundary geometry affects $L^q$ approximation by shallow ReLU neural networks on convex bodies $K
Prediction models degrade when the deployment population differs from the one used to train them, and training
We introduce the Axiomatic Learning Structures (ALS) and Axiomatic Probabilistic Structure (APS) in order to s
Digital SRAM compute-in-memory (CIM) macros enable energy-efficient CNN inference by eliminating off-chip weig
Traffic forecasting requires modeling complex temporal dynamics and the spatial dependencies imposed by road t
Online resource allocation faces two coupled challenges: (1) allocating scarce resources efficiently under unc
Fine-tuning Large Language Models (LLMs) on instruction data can quietly erode their safety, even when no trai
Metis is a bet that better data, better training methods and new architectures can buy what a bigger parameter
Diffusion vision-language models (DVLMs) process their entire input at every denoising step. Since visual toke
Spiking neural networks (SNNs) encode information not only in whether neurons fire but also in when they fire,
LLMs now prove hard theorems in Lean. Can they also design algorithms and prove their correctness and efficien
Many decision-making tasks are inherently multi-objective. Interactive evolutionary multi-objective optimizati
Hyperdimensional computing (HDC) offers efficient learning and inference for resource-constrained systems, yet
We study how learning algorithm design shapes equilibrium selection in infinite-horizon discounted Markov game
We study the computational complexity of clairvoyant strategy optimization against a known online learner in f
Improving on a baseline requires exploring away from it, precisely where its performance may be lost. We propo
The Hylland-Zeckhauser (HZ) mechanism provides a market-based approach to one-sided matching, making efficient
Direct-policy World Action Models (WAM) have demonstrated the benefits of decoupling action generation from ex
kserveは、分散型生成的および予測性AI推論プラットフォームであり、可 scale, multi-framework デプロイをサポートして Kubernetes 上に展開されることができます。
A curated collection of papers, benchmarks, surveys, and tools for model quantization, covering low-bit networ
Text injection trains speech recognition models on unpaired text alongside paired audio. We feed the speech en
Within the broader landscape of artificial intelligence architectures, this paper introduces a compact, edge-c
Time-sensitive neuromorphic architectures require predictable interconnects to preserve precise event timing.
Developers download quantized large language models (LLMs) from public registries to run them locally. None of
Language agents increasingly act through external tools and interactive environments, making the interface bet
Reservoir computing shifts the focus from training a system's internal dynamics to designing dynamics that exp
GUI grounding requires locating small controls in cluttered screenshots and identifying which control a natura
この論文では、Causal-Forcing: Autoregressive Diffusion Distillation Done Right for High-Quality Real-Time Interactive
Hardware-aware neural architecture search (NAS) requires jointly optimizing predictive accuracy and device-spe
Event-based spiking neural networks (SNNs) encode information through temporal firing dynamics, but the same s
Spiking neural networks (SNNs) are well suited to event-driven edge intelligence, but the arithmetic complexit
In AI-assisted service systems, the server can choose how much processing effort to spend on each job, which w
Streaming audio-visual target speaker extraction (AVTSE) is essential for latency-sensitive applications. Howe
《李宏毅深度学习教程》（李宏毅老师推荐👍，苹果书🍎），PDF下载地址：https://github.com/datawhalechina/leedl-tutorial/releases
In neural networks, continual learning (CL) is a prominent approach for enabling deployed systems to adapt to
Continual learning (CL) in neural networks is a promising learning approach for updating the knowledge of depl
Probabilistic computing has proved useful for combinatorial optimization, but conventional implementations rel
Generative AI assistants are becoming an advertising medium, embedding sponsored content in their answers. Thi
We study finite many-to-one matching markets with indifferences on both sides and publicly known firm preferen
Iterative combinatorial auctions are used in spectrum allocation, transportation, and industrial procurement,
Low-light image enhancement (LLIE) aims to recover well-exposed and perceptually faithful images from severely
Harness optimization, which improves an agent's prompts, tools, and workflows, is an increasingly effective wa
Spiking Neural Networks (SNNs) promise energy-efficient inference through sparse, event-driven computation, ye
EEG foundation models are advancing at an unprecedented pace, yet their evaluation is driven by downstream acc
Recent electroencephalography (EEG) foundation models increasingly emphasize model scale, potentially limiting
Dropout and zoneout are routinely added to long short-term memory (LSTM) and gated recurrent unit (GRU) networ
Bergemann, Brooks, and Morris (2015) characterize the consumer and producer surplus pairs that segmentation ca
We give a bounded envy-free cake-cutting protocol for any number $n$ of agents in the Robertson--Webb model, w
An $n$-player game in normal form is said to be of format $\boldsymbol{d} = (d_1, \dots, d_n)$ if the $i$-th p
Edge medical segmentation must reconcile small model size, high accuracy, and fast inference. TernScope addres
Clinical registries and wearable sensors support early health-risk assessment, yet models trained on these rec
Smart-city applications often combine scarce labeled data with abundant unlabeled observations. Semi-supervise
Envy-free cake cutting asks how to divide a resource so that no agent prefers another agent's share. Finding s
Robotic systems with revolute joints evolve on circular state spaces, yet most control architectures represent
Trust and reputation mechanisms can reduce the influence of malicious clients in federated learning (FL), but
The goal of this paper is to distill insight around the trade-offs present on Bayesian equilibria of transacti
The metric distortion framework asks voting rules to elect a candidate close to the voters using only their ra
Realistic market impact is essential for generative limit order book (LOB) models to serve as world models for
Agents that run for a long time have a memory problem, and it is not the one most systems optimize for. Storin
How can a shared resource be allocated without money when agents have private, stochastic demands that can be
AwesomeOPDはオンPolicy distillationの最適化用リストである。オンPolicy distillationでは、学習済みモデルを小さくすることでモデルを高速化する。
We consider the metric distortion framework for voting, where $n$ voters and $m$ candidates are embedded in an
Open-ended search combined with foundation models has produced advances such as faster algorithms, new results
Universal function approximation underpins neural networks' capacity to represent nonlinear relationships, but
On neuromorphic chips, spiking neural networks (SNNs) consume far less energy than artificial neural networks
Pacing equilibria describe how budget-constrained advertisers clear an ad market, but computing one at product
The United States allocates representatives to each state in proportion to its population, while many parliame
We study the inverse problem of recovering unknown parameters governing the dynamics and rewards of Mean Field
AI services are prone to moral hazard, as model providers can silently route requests to cheaper, lower-qualit
Blockchain forensics is adversarial and open-ended: money launderers split, swap, and bridge funds, cut their
この研究では、ストリーミングでのビデオ超解像化を実現する一ステップのディフュージョンフレームワーク「FlashVSR」を提案しています。このフレームワークは、局所制限の疎注意と小さい条件的デコーダを組み合わせて、効率的に
The rapid growth in the size of state-of-the-art machine-learning (ML) models has led to increasing compute de
Our understanding of turn-based stochastic games is reminiscent of our understanding of linear programming in
Repeated Bertrand competition is a stringent benchmark for algorithmic pricing: in homogeneous-goods markets,
Spoken Dialogue Models (SDMs) typically share a speech-token space for generation and synthesis, leading to lo
Closed-loop neuromodulation requires sleep staging that respects stage-specific minimum dwells: stimulation po
Opponent histories can mislead after behavioral shifts. SafeAdapt-Poker formulates budgeted adaptation under o
We study proportional cake cutting under an equal-length constraint, where each of $n$ agents must receive a b
Proof-of-Stake (PoS) systems offer a scalable and energy-efficient alternative to Proof-of-Work (PoW). The eco
The Continuous Thought Machine (CTM) is a recurrent network that attends to inputs and makes predictions based
We show that two Transformers representing the same function can follow different training trajectories. For e
Spiking Neural Networks (SNNs) are promising for energy-efficient neuromorphic computing, but they still suffe
Modern Transformers rely on many scale-control mechanisms, including normalization, residual scaling, attentio
Simple aggregation of independent responses ("wisdom of crowds") generally improves collective performance thr
Assigning routine tasks to cheaper large language model (LLM) providers can reduce generation cost, but switch
Regret matching (and its variants) and counterfactual regret minimization are central algorithms in practice f
Deep reinforcement learning agents in partially observable competitive games must act on beliefs about an unse
Plurality-voting ensembles aggregate the predictions of multiple base classifiers and output the most voted cl
We study a problem that unifies two major generalizations of popular matchings in the one-sided preference mod
Although spiking neural networks (SNNs) show promise in energy-efficient computing through event-driven comput
Online motor imagery-based brain-computer interfaces (MI-BCIs) typically decode continuous EEG using fixed-len
Architecture attribute predictors reduce the cost of neural architecture search by estimating accuracy or late
We present an overview of R package called \texttt{naspaclust}, a tool for spatial fuzzy clustering improvemen
We prove that, for every fixed rational $c>0$, computing a $2^{-n^c}$-approximate Nash equilibrium in an $n$-p
Next-generation wireless networks are expected to support high data rates and reliable connectivity while meet
🔥 On-Policy Self-Distillation in Diffusion Models
Small Adam-trained shared-representation networks can reach equivalent selected mean readout measurements yet
Functional magnetic resonance imaging (fMRI) commonly characterizes large-scale brain organization through fun
Efficient neural architecture search (NAS) requires accurate performance estimation under limited evaluation b
Spiking neural networks target energy-efficient language modeling through sparse, binary activations, but the
Latent recursion repeatedly processes internal representations through the same neural network. Recent work ha
Regularization-based continual learning protects each parameter in proportion to an estimate of its importance
Clinical sleep medicine repeatedly documents profound divergences between objective sleep duration and subject
We study when communication emerges between autonomous agents and what role an innate action selection archite
画面の生成モデルであるHunyuanVideoを開発した。HunyuanVideoは、複雑なシーケンスを生成する能力を持つ。
Accurate forecasting models are usually large, expensive to update online, and fixed in architecture once trai
Biological connectomes encode structured solutions to visual computation that may provide reusable inductive b
Where generalization capacity--the ability to extract context-invariant relational structures--first emerges r
Sensing-aided beam prediction can reduce the beam-training overhead of millimeter-wave (mmWave) systems by lev
Time series forecasting models are typically compared on pointwise error, which scores a prediction in isolati
GUI for a Vocal Remover that uses Deep Neural Networks.
Neuromorphic hardware combines memory (synapses) and computation (neurons) into the same silicon substrate to
Swarm robotics presents a robust and cost-effective paradigm for advanced automation in complex, dynamic envir
Landscape analysis has been successfully applied to understand complex optimisation problems, gain insights in
Living neuronal networks transform inputs through recurrent cellular and population dynamics, yet it is unknow
Frameworks that use large language models for scientific discovery typically rely on a fixed, human-designed a
LLM-guided evolutionary methods, such as AlphaEvolve, have emerged as powerful approaches for challenging comp
Integer-valued leaky integrate-and-fire (I-LIF) neurons and spike firing approximation (SFA) reduce temporal t
The investment acceleration principle is a heuristic for modelling investment time series out of consumption t
Deploying a pretrained controller exposes it to conditions that are absent from its training data. Sensor drif
Hybrid quantum classical neural networks integrate parameterized quantum circuits (PQCs) with established deep
Aggregate accuracy cannot reveal which utterances are locally vulnerable or how internal activity changes when
Edge intelligence requires models to sense continuously in real time and to keep adapting on-device, all under
This paper presents a construction of a small universal machine within the framework of Multi-set Reaction Sys
Spiking Neural Networks (SNNs) have attracted increasing attention due to their impressive temporal dynamics,
Low-cost NAS must both explore high-performing architectures and identify them reliably, yet reducing evaluati
End-to-end backpropagation has been the dominant mode of training in deep learning, allowing for the coordinat
Pseudo-random number generation often requires trade-offs among quality, power consumption, and bandwidth to p
Spiking neural networks (SNNs) offer low-energy sequence modeling through sparse, event-driven computation. Ho
Compact regulatory DNA can free up space in vector payloads, reduce synthesis and assay burden, and expose whi
Randomised copies of a connectome are the usual baseline for asking whether measured wiring matters, and the a
Parameter-efficient reinforcement learning aims to improve reasoning with a compact trainable interface to a p
Large language models are increasingly participating in complex real-world tasks in the form of algorithm-desi
This work evaluates the direct transfer of a co-evolved communication protocol from a 2D simulation to a 3D ph
Hyperdimensional computing and vector symbolic architectures often represent quantized scalar levels by dense
Every day, people continuously infer situational context and adjust the way they understand and remember the w
We study whether the coordinate dependence of the improved Self-Organizing Migrating Algorithm (iSOMA) can be
We propose a new spiking neural network (SNN) design to process static images and event streams using time-to-
Clinical scalp electroencephalography (EEG) offers a noninvasive window into neural dynamics of neuropsychiatr
Hardware accelerators based on physical dynamical systems offer an attractive route toward energy-efficient re
Adam and its variants dominate neural network training, but a single run only reveals whether a configuration
A critical analysis of contemporary approaches to the study of conscious states. The review focuses on methods
Sensor networks and radars can deliver signals as organized sets, e.g. ordered signals, signals describing ran
We study the knowledge matrix of a trained feedforward network as a higher representation of its inputs. A net
A central aim of neuromorphic computing is to provide a viable alternative to highly energy-intensive Transfor