paper-with-me

Papers

JEPA-Reasoner: Decoupling Latent Reasoning from Token Generation

2025-12-22 · Bingyang Kelvin Liu, Ziyu Patrick Chen, David P. Woodruff arxiv

Current autoregressive language models couple high-level reasoning and low-level token generation into a single sequential process, making the reasoning trajectory vulnerable to compounding expression errors. We propose JEPA-Reasoner, a novel architectural paradigm that decouples these tasks using a Joint-Embedding Predictive Architecture (JEPA) for pure latent-space reasoning and a separate Talker module for linguistic reconstruction. By isolating the reasoning engine from the discrete token-sampling process, our architecture enables: (1) Error Containment, where token-level failures cannot propagate into the latent reasoning chain; (2) Continuous Guidance, providing the generator with access to the entire lossless reasoning trajectory; and (3) Representation of Uncertainty, allowing the model to maintain multiple hypotheses via mixed latent vectors. Controlled experiments on synthetic and natural language tasks demonstrate that this decoupling enables a 0.9B model to achieve a 149.5\% improvement in 8-shot GSM8K accuracy over a coupled Transformer baseline trained on identical data. This work shifts the focus from scaling coupled models to investigating decoupled architectures as a more robust foundation for complex reasoning.

📄 PDF Abstract BibTeX arXiv:2512.19171

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Discrete JEPA: Learning Discrete Token Representations without Reconstruction

2025-06-17 · Junyeob Baek, Hosung Lee, Christopher Hoang, Mengye Ren 외

The cornerstone of cognitive intelligence lies in extracting hidden patterns from observations and leveraging these principles to systematically predict future outcomes. However, current image tokenization methods demons…

Logical Reasoning

JEPA-DNA: Grounding Genomic Foundation Models through Joint-Embedding Predictive Architectures

2026-02-19 · Ariel Larey, Elay Dahan, Amit Bleiweiss, Raizy Kellerman 외 arxiv

Genomic Foundation Models (GFMs) typically rely on Masked Language Modeling (MLM) or Next-Token Prediction (NTP) to learn the "Laws of Nature". While effective at capturing local syntax, these generative paradigms priori…

SLPO: Scaling Latent Reasoning via a Surrogate Policy

2026-07-22 · Runyang You, Zhiyuan Liu, Yongqi Li, Wenjie Li arxiv

Reinforcement learning with verifiable rewards has become the predominant recipe for eliciting test-time scaling in explicit Chain-of-Thought reasoners. Yet this scaling path remains computationally costly, since every i…

Reinforcement Learning

Unified Embodied VLM Reasoning with Robotic Action via Autoregressive Discretized Pre-training

2025-12-30 · Yi Liu, Sukai Wang, Dafeng Wei, Xiaowei Cai 외 arxiv

General-purpose robotic systems operating in open-world environments must achieve both broad generalization and high-precision action execution, a combination that remains challenging for existing Vision-Language-Action …

Continuous Control

HQ-JEPA: Hybrid Quantum Joint-Embedding Predictive Architecture for Cross-Modal Remote Sensing Representation Learning

2026-05-29 · Md Aminur Hossain, Ayush V. Patel, Sanjay K. Singh, Biplab Banerjee arxiv

We introduce HQ-JEPA, a hybrid quantum-classical joint-embedding predictive architecture for cross-modal remote sensing representation learning. The proposed framework extends JEPA-style masked latent prediction to paire…

Representation Learning