Reasoning as Energy Minimization over Structured Latent Trajectories

• URL: http://arxiv.org/abs/2603.28248v1
• Date: Mon, 30 Mar 2026 10:20:56 GMT
• Title: Reasoning as Energy Minimization over Structured Latent Trajectories
• Abstract summary: We propose Energy-Based Reasoning via Structured Latent Planning (EBRM)<n>EBRM models optimization of a multi-step latent trajectory $z_1:T$ under a learned energy function $E(h_x, z)$.<n>Training combines supervised encoder-decoder learning with contrastive energy shaping using hard negatives.<n>We analyze this behavior through per-step decoding, latent drift tracking, and gradient decomposition.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Single-shot neural decoders commit to answers without iterative refinement, while chain-of-thought methods introduce discrete intermediate steps but lack a scalar measure of reasoning progress. We propose Energy-Based Reasoning via Structured Latent Planning (EBRM), which models reasoning as gradient-based optimization of a multi-step latent trajectory $z_{1:T}$ under a learned energy function $E(h_x, z)$. The energy decomposes into per-step compatibility, transition consistency, and trajectory smoothness terms. Training combines supervised encoder-decoder learning with contrastive energy shaping using hard negatives, while inference performs gradient descent or Langevin dynamics over $z$ and decodes from $z_T$. We identify a critical failure mode: on CNF logic satisfaction, latent planning reduces accuracy from $\approx 95\%$ to $\approx 56\%$. This degradation arises from a distribution mismatch, where the decoder is trained on encoder outputs $h_x$ but evaluated on planner outputs $z_T$ that drift into unseen latent regions. We analyze this behavior through per-step decoding, latent drift tracking, and gradient decomposition. To address it, we propose dual-path decoder training and latent anchoring. We further introduce a six-part ablation protocol covering component contributions, trajectory length, planner dynamics, initialization, decoder training distribution, and anchor weight. Experiments on three synthetic tasks show that energy decreases monotonically and induces structured latent trajectories on graph and logic tasks, while remaining flat on arithmetic ($r = 0.073$), indicating a negative result. Code is available at https://github.com/dkjo8/ebr-via-structured-latent-planning.

Related papers

• Think Wider: Mitigating Latent Rank Collapse in Implicit Chain-of-Thought Reasoning [3.216436532980238]
  Chain-of-thought (CoT) reasoning improves the reasoning ability of large language models by introducing intermediate computation.<n>However, explicit rationales increase decoding length, latency, and context cost.<n>Implicit CoT offers a more efficient alternative by moving intermediate reasoning into continuous latent states.
  arXiv  Detail & Related papers  (2026-09-07T12:18:51Z)
• Imbuing Large Language Models with Bidirectional Logic for Robust Chain Repair [44.80087038178069]
  We introduce Teleological Reasoning Infilling (TRI), a training framework that endows decoder-only transformers with a native emphgoal-conditioned bridging capability.<n>At inference, TRI operates as a surgical repair module within a dual-system loop.<n>Experiments on three benchmarks demonstrate that TRI achieves state-of-the-art performance across all tasks, while reducing per-problem token expenditure by 31.2%.
  arXiv  Detail & Related papers  (2026-06-03T15:58:48Z)
• Transformers Provably Learn to Internalize Chain-of-Thought [65.41010769606844]
  Implicit Chain-of-Thought (ICoT) trains models to internalize intermediate steps within their hidden states.<n>Log-ICoT removes thinking tokens one at a time, reducing the number of stages from linear in $k$ to logarithmic.<n>Experiments on multi-layer transformers confirm the theory and visualize how reasoning is progressively absorbed into deeper layers.
  arXiv  Detail & Related papers  (2026-05-27T15:17:06Z)
• $\ abla$-Reasoner: LLM Reasoning via Test-Time Gradient Descent in Latent Space [71.23672814629448]
  $nabla$-Reasoner is an iterative generation framework that integrates differentiable optimization over token logits into the decoding loop.<n>$nabla$-Reasoner achieves over 20% accuracy improvement on a challenging mathematical reasoning benchmark.
  arXiv  Detail & Related papers  (2026-03-05T08:42:54Z)
• Operator Learning Using Weak Supervision from Walk-on-Spheres [81.26322147849918]
  Training neural PDE solvers is often bottlenecked by expensive data generation or unstable physics-informed neural network (PINN)<n>We propose an alternative approach using Monte Carlo approaches to estimate the solution to the PDE as a process for weak supervision during training.
  arXiv  Detail & Related papers  (2026-03-01T17:23:39Z)
• Unbiased Gradient Estimation for Event Binning via Functional Backpropagation [64.88399635309918]
  We propose a novel framework for unbiased gradient estimation of arbitrary binning functions by synthesizing weak derivatives during backpropagation.<n>We achieve 9.4% lower EPE in self-supervised optical flow, and 5.1% lower RMS error in SLAM, demonstrating broad benefits for event-based visual perception.
  arXiv  Detail & Related papers  (2026-02-13T04:05:03Z)
• Gradient Descent as a Perceptron Algorithm: Understanding Dynamics and Implicit Acceleration [67.12978375116599]
  We show that the steps of gradient descent (GD) reduce to those of generalized perceptron algorithms.<n>This helps explain the optimization dynamics and the implicit acceleration phenomenon observed in neural networks.
  arXiv  Detail & Related papers  (2025-12-12T14:16:35Z)
• Renormalizable Spectral-Shell Dynamics as the Origin of Neural Scaling Laws [2.779943773196378]
  We show that deep-network training obeys a simple macroscopic structure despite highly nonlinear optimization dynamics.<n>For mean-squared error loss, the training error evolves as $dot e_t=-M(t)e_t$ with $M(t)=J_(t)J_(t)!*$, a time-dependent self-adjoint operator induced by the network Jacobian.<n>This framework explains neural scaling laws and double descent, and unifies lazy (NTK-like) training and feature learning as two limits of the same spectral-shell
  arXiv  Detail & Related papers  (2025-12-11T08:38:46Z)
• INC: An Indirect Neural Corrector for Auto-Regressive Hybrid PDE Solvers [61.84396402100827]
  We propose the Indirect Neural Corrector ($mathrmINC$), which integrates learned corrections into the governing equations.<n>$mathrmINC$ reduces the error amplification on the order of $t-1 + L$, where $t$ is the timestep and $L$ the Lipschitz constant.<n>We test $mathrmINC$ in extensive benchmarks, covering numerous differentiable solvers, neural backbones, and test cases ranging from a 1D chaotic system to 3D turbulence.
  arXiv  Detail & Related papers  (2025-11-16T20:14:28Z)
• Tortoise and Hare Guidance: Accelerating Diffusion Model Inference with Multirate Integration [5.634791333443662]
  Tortoise and Hare Guidance (THG) is a training-free strategy that accelerates diffusion sampling while maintaining high-fidelity generation.<n>We show that THG reduces the number of function evaluations (NFE) by up to 30% with virtually no loss in generation fidelity.<n>Our findings highlight the potential of multirate formulations for diffusion solvers, paving the way for real-time high-quality image synthesis without any model retraining.
  arXiv  Detail & Related papers  (2025-11-06T07:08:58Z)
• Decentralized Nonconvex Composite Federated Learning with Gradient Tracking and Momentum [78.27945336558987]
  Decentralized server (DFL) eliminates reliance on client-client architecture.<n>Non-smooth regularization is often incorporated into machine learning tasks.<n>We propose a novel novel DNCFL algorithm to solve these problems.
  arXiv  Detail & Related papers  (2025-04-17T08:32:25Z)
• BP(\lambda): Online Learning via Synthetic Gradients [6.581214715240991]
  Training recurrent neural networks typically relies on backpropagation through time (BPTT) In their implementation synthetic gradients are learned through a mixture of backpropagated gradients and bootstrapped synthetic gradients. Inspired by the accumulate $mathrmTD(lambda)$ in RL, we propose a fully online method for learning synthetic gradients which avoids the use of BPTT altogether.
  arXiv  Detail & Related papers  (2024-01-13T11:13:06Z)
• Low PAPR MIMO-OFDM Design Based on Convolutional Autoencoder [20.544993155126967]
  A new framework for peak-to-average power ratio ($mathsfPAPR$) reduction and waveform design is presented. A convolutional-autoencoder ($mathsfCAE$) architecture is presented. We show that a single trained model covers the tasks of $mathsfPAPR$ reduction, spectrum design, and $mathsfMIMO$ detection together over a wide range of SNR levels.
  arXiv  Detail & Related papers  (2023-01-11T11:35:10Z)
• Asynchronous Training Schemes in Distributed Learning with Time Delay [17.259708772713164]
  In the context of distributed deep learning, the issue of stale weights or gradients could result in poor algorithmic performance. In this paper, we propose a different approach to tackle the issue of stale weights or gradients. One practical variant of PC-ASGD is also proposed by adopting a condition to help with the determination of the tradeoff parameter.
  arXiv  Detail & Related papers  (2022-08-28T07:14:59Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.