Gradient Free Deep Reinforcement Learning With TabPFN

• URL: http://arxiv.org/abs/2509.11259v1
• Date: Sun, 14 Sep 2025 13:09:58 GMT
• Title: Gradient Free Deep Reinforcement Learning With TabPFN
• Abstract summary: We propose TabPFN RL, a gradient free deep RL framework that repurposes the meta trained transformer TabPFN as a Q function approximator.<n>We use TabPFN to predict Q values using inference only, thereby eliminating the need for back propagation at both training and inference.<n>Our results establish prior fitted networks such as TabPFN as a viable foundation for fast and computationally efficient RL, opening new directions for gradient free RL with large pre trained transformers.
• Score: 23.140580395364207
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Gradient based optimization is fundamental to most modern deep reinforcement learning algorithms, however, it introduces significant sensitivity to hyperparameters, unstable training dynamics, and high computational costs. We propose TabPFN RL, a novel gradient free deep RL framework that repurposes the meta trained transformer TabPFN as a Q function approximator. Originally developed for tabular classification, TabPFN is a transformer pre trained on millions of synthetic datasets to perform inference on new unseen datasets via in context learning. Given an in context dataset of sample label pairs and new unlabeled data, it predicts the most likely labels in a single forward pass, without gradient updates or task specific fine tuning. We use TabPFN to predict Q values using inference only, thereby eliminating the need for back propagation at both training and inference. To cope with the model's fixed context budget, we design a high reward episode gate that retains only the top 5% of trajectories. Empirical evaluations on the Gymnasium classic control suite demonstrate that TabPFN RL matches or surpasses Deep Q Network on CartPole v1, MountainCar v0, and Acrobot v1, without applying gradient descent or any extensive hyperparameter tuning. We discuss the theoretical aspects of how bootstrapped targets and non stationary visitation distributions violate the independence assumptions encoded in TabPFN's prior, yet the model retains a surprising generalization capacity. We further formalize the intrinsic context size limit of in context RL algorithms and propose principled truncation strategies that enable continual learning when the context is full. Our results establish prior fitted networks such as TabPFN as a viable foundation for fast and computationally efficient RL, opening new directions for gradient free RL with large pre trained transformers.

Related papers

• REVES: REvision and VErification--Augmented Training for Test-Time Scaling [53.197756110943395]
  We propose a two-stage iterative framework that alternates between online data/prompt augmentation and policy optimization.<n>We observe gains of +6.5 points over the RL baseline and +4.0 points over standard multi-turn training.
  arXiv  Detail & Related papers  (2026-06-17T10:37:23Z)
• PriFT: Prior-Support Guided Supervised Fine-Tuning [74.65198014829393]
  Supervised fine-tuning (SFT) is an efficient approach for downstream task adaptation.<n>SFT fits fixed demonstrations token by token, including targets poorly aligned with the model's pretrained distribution.<n>We propose PriFT, which derives token weights from a frozen pretrained reference to obtain a stable reweighting signal unaffected by fine-tuning.
  arXiv  Detail & Related papers  (2026-06-08T12:14:06Z)
• Mitigating Label Shift in Tabular In-Context Learning via Test-Time Posterior Adjustment [12.754011132549167]
  DistPFN is the first test-time posterior adjustment method designed for tabular foundation models.<n>We evaluate our methods on over 250 OpenML datasets, demonstrating substantial improvements for various TabPFN-based models.
  arXiv  Detail & Related papers  (2026-05-06T00:01:47Z)
• Low-rank Optimization Trajectories Modeling for LLM RLVR Acceleration [88.7332439088792]
  scaling reinforcement learning with verifiable rewards (RLVR) for large language models (LLMs) has emerged as an effective training paradigm.<n>We propose the textbfNonlinear textbfExtrapolation of low-rank trajectories (textbfNExt)<n>Our method reduces computational overhead by approximately 37.5% while remaining compatible with a wide range of RLVR algorithms and tasks.
  arXiv  Detail & Related papers  (2026-04-13T13:28:12Z)
• TabICLv2: A better, faster, scalable, and open tabular foundation model [18.594859017648346]
  We introduce TabICLv2, a new state-of-the-art foundation model for regression and classification built on three pillars.<n>Tabiclv2 generalizes effectively to million-scale datasets under 50GB GPU memory while being markedly faster than RealTabPFN-2.5.
  arXiv  Detail & Related papers  (2026-02-11T18:51:02Z)
• Learning in Context, Guided by Choice: A Reward-Free Paradigm for Reinforcement Learning with Transformers [55.33468902405567]
  We propose a new learning paradigm, In-Context Preference-based Reinforcement Learning (ICPRL), in which both pretraining and deployment rely solely on preference feedback.<n>ICPRL enables strong in-context generalization to unseen tasks, achieving performance comparable to ICRL methods trained with full reward supervision.
  arXiv  Detail & Related papers  (2026-02-09T03:42:16Z)
• The Path Not Taken: RLVR Provably Learns Off the Principals [85.41043469428365]
  We show that sparsity is a surface artifact of a model-conditioned optimization bias.<n>We mechanistically explain these dynamics with a Three-Gate Theory.<n>We provide a parameter-level characterization of RLVR's learning dynamics.
  arXiv  Detail & Related papers  (2025-11-11T18:49:45Z)
• Q-SFT: Q-Learning for Language Models via Supervised Fine-Tuning [62.984693936073974]
  Value-based reinforcement learning can learn effective policies for a wide range of multi-turn problems.<n>Current value-based RL methods have proven particularly challenging to scale to the setting of large language models.<n>We propose a novel offline RL algorithm that addresses these drawbacks, casting Q-learning as a modified supervised fine-tuning problem.
  arXiv  Detail & Related papers  (2024-11-07T21:36:52Z)
• PACE: Marrying generalization in PArameter-efficient fine-tuning with Consistency rEgularization [35.922096876707975]
  PACE is a generalization of PArameter-efficient fine-tuning with Consistency rEgularization.<n>It implicitly regularizes gradients for enhanced generalization, but also implicitly aligns the fine-tuned and pre-trained models to retain knowledge.<n>It also improves LoRA in text classification (GLUE) and mathematical reasoning.
  arXiv  Detail & Related papers  (2024-09-25T17:56:00Z)
• Just How Flexible are Neural Networks in Practice? [89.80474583606242]
  It is widely believed that a neural network can fit a training set containing at least as many samples as it has parameters. In practice, however, we only find solutions via our training procedure, including the gradient and regularizers, limiting flexibility.
  arXiv  Detail & Related papers  (2024-06-17T12:24:45Z)
• Tokenize features, enhancing tables: the FT-TABPFN model for tabular classification [13.481699494376809]
  FT-TabPFN is an enhanced version of TabPFN that includes a novel Feature Tokenization layer to better handle classification features. Our full source code is available for community use and development.
  arXiv  Detail & Related papers  (2024-06-11T02:13:46Z)
• On Mesa-Optimization in Autoregressively Trained Transformers: Emergence and Capability [34.43255978863601]
  Several suggest that transformers learn a mesa-optimizer during autorere training. We show that a stronger assumption related to the moments of data is the sufficient necessary condition that the learned mesa-optimizer can perform.
  arXiv  Detail & Related papers  (2024-05-27T05:41:06Z)
• TuneTables: Context Optimization for Scalable Prior-Data Fitted Networks [90.00817095558094]
  Prior-data fitted networks (PFNs) make use of pretraining and in-context learning to achieve strong performance on new tasks in a single forward pass. We introduce TuneTables, a parameter-efficient fine-tuning strategy for PFNs that compresses large datasets into a smaller learned context. We show that TuneTables can be used as an interpretability tool and can even be used to mitigate biases by optimizing a fairness objective.
  arXiv  Detail & Related papers  (2024-02-17T00:02:23Z)
• Sparse is Enough in Fine-tuning Pre-trained Large Language Models [98.46493578509039]
  We propose a gradient-based sparse fine-tuning algorithm, named Sparse Increment Fine-Tuning (SIFT) We validate its effectiveness on a range of tasks including the GLUE Benchmark and Instruction-tuning.
  arXiv  Detail & Related papers  (2023-12-19T06:06:30Z)
• Winner-Take-All Column Row Sampling for Memory Efficient Adaptation of Language Model [89.8764435351222]
  We propose a new family of unbiased estimators called WTA-CRS, for matrix production with reduced variance. Our work provides both theoretical and experimental evidence that, in the context of tuning transformers, our proposed estimators exhibit lower variance compared to existing ones.
  arXiv  Detail & Related papers  (2023-05-24T15:52:08Z)
• Scalable Real-Time Recurrent Learning Using Columnar-Constructive Networks [19.248060562241296]
  We propose two constraints that make real-time recurrent learning scalable. We show that by either decomposing the network into independent modules or learning the network in stages, we can make RTRL scale linearly with the number of parameters. We demonstrate the effectiveness of our approach over Truncated-BPTT on a prediction benchmark inspired by animal learning and by doing policy evaluation of pre-trained policies for Atari 2600 games.
  arXiv  Detail & Related papers  (2023-01-20T23:17:48Z)
• ORFit: One-Pass Learning via Bridging Orthogonal Gradient Descent and Recursive Least-Squares [5.430441358049335]
  We investigate the problem of one-pass learning, in which a model is trained on sequentially arriving data without retraining on previous datapoints.<n>We propose Orthogonal Recursive Fitting (ORFit), an algorithm for one-pass learning which seeks to perfectly fit each new datapoint while minimally altering the predictions on previous datapoints.
  arXiv  Detail & Related papers  (2022-07-28T02:01:31Z)
• Large-Scale Gradient-Free Deep Learning with Recursive Local Representation Alignment [84.57874289554839]
  Training deep neural networks on large-scale datasets requires significant hardware resources. Backpropagation, the workhorse for training these networks, is an inherently sequential process that is difficult to parallelize. We propose a neuro-biologically-plausible alternative to backprop that can be used to train deep networks.
  arXiv  Detail & Related papers  (2020-02-10T16:20:02Z)

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.