Actor-Critic based Online Data Mixing For Language Model Pre-Training

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Ma, Jing, Dang, Chenhao, Liao, Mingjie
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Bridging the Gap: Enabling Soft Actor Critic for High Performance Legged Locomotion
di: Sabatini, Gianluca, et al.
Pubblicazione: (2026)

SMAC: Score-Matched Actor-Critics for Robust Offline-to-Online Transfer
di: de Lara, Nathan Samuel, et al.
Pubblicazione: (2026)

Collaborative Yet Personalized Policy Training: Single-Timescale Federated Actor-Critic
di: Wang, Leo Muxing, et al.
Pubblicazione: (2026)

SAC-GLAM: Improving Online RL for LLM agents with Soft Actor-Critic and Hindsight Relabeling
di: Gaven, Loris, et al.
Pubblicazione: (2024)

Average-Reward Soft Actor-Critic
di: Adamczyk, Jacob, et al.
Pubblicazione: (2025)

Value Improved Actor Critic Algorithms
di: Oren, Yaniv, et al.
Pubblicazione: (2024)

Diffusion Actor-Critic with Entropy Regulator
di: Wang, Yinuo, et al.
Pubblicazione: (2024)

Revisiting Discrete Soft Actor-Critic
di: Zhou, Haibin, et al.
Pubblicazione: (2022)

DSAC: Distributional Soft Actor-Critic for Risk-Sensitive Reinforcement Learning
di: Ma, Xiaoteng, et al.
Pubblicazione: (2020)

Distributional Soft Actor-Critic with Diffusion Policy
di: Liu, Tong, et al.
Pubblicazione: (2025)

Flow Actor-Critic for Offline Reinforcement Learning
di: Chae, Jongseong, et al.
Pubblicazione: (2026)

FM-EAC: Feature Model-based Enhanced Actor-Critic for Multi-Task Control in Dynamic Environments
di: Zhou, Quanxi, et al.
Pubblicazione: (2025)

Causal Policy Learning in Reinforcement Learning: Backdoor-Adjusted Soft Actor-Critic
di: Vo, Thanh Vinh, et al.
Pubblicazione: (2025)

Contrastive Language-Image Pre-Training Model based Semantic Communication Performance Optimization
di: Yang, Shaoran, et al.
Pubblicazione: (2025)

Lightweight Geometric Adaptation for Training Physics-Informed Neural Networks
di: An, Kang, et al.
Pubblicazione: (2026)

AutoScale: Scale-Aware Data Mixing for Pre-Training LLMs
di: Kang, Feiyang, et al.
Pubblicazione: (2024)

Relational Object-Centric Actor-Critic
di: Ugadiarov, Leonid, et al.
Pubblicazione: (2023)

Offline Actor-Critic Reinforcement Learning Scales to Large Models
di: Springenberg, Jost Tobias, et al.
Pubblicazione: (2024)

SACn: Soft Actor-Critic with n-step Returns
di: Łyskawa, Jakub, et al.
Pubblicazione: (2025)

Actor-Critics Can Achieve Optimal Sample Efficiency
di: Tan, Kevin, et al.
Pubblicazione: (2025)

Scalable Neighborhood-Based Multi-Agent Actor-Critic
di: Goppelsroeder, Tim, et al.
Pubblicazione: (2026)

Revisiting Mixture Policies in Entropy-Regularized Actor-Critic
di: He, Jiamin, et al.
Pubblicazione: (2026)

Cal-QL: Calibrated Offline RL Pre-Training for Efficient Online Fine-Tuning
di: Nakamoto, Mitsuhiko, et al.
Pubblicazione: (2023)

Functional Critics Are Essential for Actor-Critic: From Off-Policy Stability to Efficient Exploration
di: Bai, Qinxun, et al.
Pubblicazione: (2025)

Dissecting Discrete Soft Actor-Critic: Limitations and Principled Alternatives
di: Asad, Reza, et al.
Pubblicazione: (2025)

Finite-time Convergence Analysis of Actor-Critic with Evolving Reward
di: Hu, Rui, et al.
Pubblicazione: (2025)

Studying the Interplay Between the Actor and Critic Representations in Reinforcement Learning
di: Garcin, Samuel, et al.
Pubblicazione: (2025)

Broad Critic Deep Actor Reinforcement Learning for Continuous Control
di: Thalagala, Shiron, et al.
Pubblicazione: (2024)

Ask-AC: An Initiative Advisor-in-the-Loop Actor-Critic Framework
di: Liu, Shunyu, et al.
Pubblicazione: (2022)

Stabilizing the Q-Gradient Field for Policy Smoothness in Actor-Critic
di: Lee, Jeong Woon, et al.
Pubblicazione: (2026)

Heavy-Ball Momentum Accelerated Actor-Critic With Function Approximation
di: Dong, Yanjie, et al.
Pubblicazione: (2024)

A Survey on Time-Series Pre-Trained Models
di: Ma, Qianli, et al.
Pubblicazione: (2023)

MDGMIX: Boundary-Aware Subgraph Mixing for Multi-Domain Graph Pre-Training
di: Zheng, Ziyu, et al.
Pubblicazione: (2026)

Mini-batch Coresets for Memory-efficient Language Model Training on Data Mixtures
di: Nguyen, Dang, et al.
Pubblicazione: (2024)

BiMix: A Bivariate Data Mixing Law for Language Model Pretraining
di: Ge, Ce, et al.
Pubblicazione: (2024)

Enabling Off-Policy Imitation Learning with Deep Actor Critic Stabilization
di: Sen, Sayambhu, et al.
Pubblicazione: (2025)

Off-Policy Actor-Critic for Adversarial Observation Robustness: Virtual Alternative Training via Symmetric Policy Evaluation
di: Nakanishi, Kosuke, et al.
Pubblicazione: (2025)

Limits of Actor-Critic Algorithms for Decision Tree Policies Learning in IBMDPs
di: Kohler, Hector, et al.
Pubblicazione: (2023)

Maximum Entropy On-Policy Actor-Critic via Entropy Advantage Estimation
di: Choe, Jean Seong Bjorn, et al.
Pubblicazione: (2024)

if (filemtime($filename) + $maxlifetime < time()) { unlink($filename); $count++; } } return $count; } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ protected function saveSession($sessId, $data): bool { $sessFile = $this->path . '/sess_' . $sessId; if ($handle = fopen($sessFile, 'w')) { $return = false; // Lock the file for exclusive access to avoid issues with multiple // processes writing session simultaneously: if (flock($handle, LOCK_EX)) { $return = fwrite($handle, $data); // Make sure that there's no trailing data by truncating the file to // the correct length: ftruncate($handle, strlen($data)); fflush($handle); flock($handle, LOCK_UN); } fclose($handle); if ($return !== false) { return true; } } // If we got this far, something went wrong with the file output... // It is tempting to throw an exception here, but this code is called // outside of the context of exception handling, so all we can do is // echo a message.

// Anecdotal testing Today and Yesterday seems to indicate destroy() // is called by the garbage collector and everything is good. // Something to keep in mind though. return 0; } /** * Write function that is called when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ public function write($sessId, $data): bool { if ($this->writesDisabled) { return true; } return $this->saveSession($sessId, $data); } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ abstract protected function saveSession($sessId, $data): bool; }

* * @return void */ public function writeClose() { // The assumption is that we're using PHP's ext/session. // session_write_close() will actually overwrite $_SESSION with an // empty array on completion -- which leads to a mismatch between what // is in the storage object and $_SESSION. To get around this, we // temporarily reset $_SESSION to an array, and then re-link it to // the storage object. // // Additionally, while you _can_ write to $_SESSION following a // session_write_close() operation, no changes made to it will be // flushed to the session handler. As such, we now mark the storage // object isImmutable. $storage = $this->getStorage(); if (! $storage->isImmutable()) { $_SESSION = $storage->toArray(true); session_write_close(); $storage->fromArray($_SESSION); $storage->markImmutable(); } } /** * Attempt to set the session name * * If the session has already been started, or if the name provided fails * validation, an exception will be raised. * * @param string $name * @return SessionManager * @throws Exception\InvalidArgumentException */ public function setName($name) { if ($this->sessionExists()) { throw new Exception\InvalidArgumentException( 'Cannot set session name after a session has already started'

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
language	"it"
VUFIND_SESSION	"86dc9905c822dd3cf58f5db4d13d4004"
ui	"standard"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790778004.9329 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "86dc9905c822dd3cf58f5db4d13d4004" ]`]
SessionState	Laminas\Stdlib\ArrayObject {#938}
FlashMessenger	Laminas\Stdlib\ArrayObject {#1433}
SessionHelper	Laminas\Stdlib\ArrayObject {#1759}

Server/Request Data

Key	Value
REDIRECT_VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
REDIRECT_HTTPS	"on"
REDIRECT_SSL_TLS_SNI	"info.umar.mx"
REDIRECT_STATUS	"200"
VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
HTTPS	"on"
SSL_TLS_SNI	"info.umar.mx"
HTTP_ACCEPT	"/"
HTTP_USER_AGENT	"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"
HTTP_COOKIE	"language=it; VUFIND_SESSION=86dc9905c822dd3cf58f5db4d13d4004; ui=standard"
HTTP_ACCEPT_ENCODING	"gzip, br, zstd, deflate"
HTTP_HOST	"info.umar.mx"
PATH	"/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/snap/bin"
SERVER_SIGNATURE	"<address>Apache/2.4.66 (Ubuntu) Server at info.umar.mx Port 443</address>\n"
SERVER_SOFTWARE	"Apache/2.4.66 (Ubuntu)"
SERVER_NAME	"info.umar.mx"
SERVER_ADDR	"200.23.222.19"
SERVER_PORT	"443"
REMOTE_ADDR	"216.73.216.147"
DOCUMENT_ROOT	"/usr/local/vufind/public"
REQUEST_SCHEME	"https"
CONTEXT_PREFIX	""
CONTEXT_DOCUMENT_ROOT	"/usr/local/vufind/public"
SERVER_ADMIN	"admin@localhost"
SCRIPT_FILENAME	"/usr/local/vufind/public/index.php"
REMOTE_PORT	"16415"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2505_23878/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2505_23878/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790778004.9278
REQUEST_TIME	1790778004

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler