When Does Tool Use Increase the Expressive Power of Finite-Precision Recurrent Models?

• URL: http://arxiv.org/abs/2607.06155v1
• Date: Tue, 07 Jul 2026 11:32:56 GMT
• Title: When Does Tool Use Increase the Expressive Power of Finite-Precision Recurrent Models?
• Abstract summary: We model any fixed finite-precision recurrent sequence model, including finite-precision state-space models (SSMs) with $B$ bits of internal state, as a deterministic finite-state controller interacting with an oracle through a finite command/observation interface.<n>A single minimal-state tool, namely a tape supporting only local $mathttread$, $mathttwrite$, and $mathttmove$ commands, makes the system Turing complete.
• Score: 22.276732262979678
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Modern sequence models are increasingly deployed as agents that interleave token generation with calls to external tools. We give an exact, architecture-level account of when such tool access increases computational expressivity. We model any fixed finite-precision recurrent sequence model, including finite-precision state-space models (SSMs) with $B$ bits of internal state, as a deterministic finite-state controller interacting with an oracle through a finite command/observation interface. Our results form a sharp dichotomy. First, tools that are themselves finite-state add essentially nothing: a product-state simulation internalizes any finite-state bounded-interface oracle with finite memory set $M$ at a cost of only $\log_2 |M| + O(1)$ additional bits, so the augmented system remains finite-state. Second, a single minimal infinite-state tool, namely a tape supporting only local $\mathtt{read}$, $\mathtt{write}$, and $\mathtt{move}$ commands, makes the system Turing complete: for every single-tape Turing machine with state set $Q$ and tape alphabet $Γ$, a controller with $O(\log |Q| + \log |Γ|)$ bits of internal memory simulates it, and we exhibit a concrete exponential separation: $\mathrm{EQ}_n$ requires $2^n$ states without tools but a single constant-size controller with the tape tool. Third, we show that this construction is realized exactly by a natural one-layer finite-precision selective affine SSM controller with binary one-hot hidden states, $\{0,1\}$ transition matrices, and zero biases. Selectivity is essential to the construction. In the supplementary material, we make all constants explicit, prove a logarithmic oracle-assisted universal simulation, where $O(\log B)$ recurrent bits suffice to simulate any $B$-state Turing machine, and prove a matching impossibility result.

Related papers

• Sharp Minimax Regret for Infinite-Memory Logistic Prediction [55.29259818039367]
  Lag $j$ can affect prediction by scale $r_j$ and enters only $n_T,j=T-j+1$ prediction rounds.<n>For every summable envelope, a localized mixture proves $cR_T(r)leq C_T(r)$.<n>For exponential and envelopes, under the stated finite-sample condition, a Toeplitz-design converse proves $cR_T(r)geq c_T(r)$.
  arXiv  Detail & Related papers  (2026-08-27T01:31:46Z)
• Quantum Coordination Advantages in AI State-Tracking Tasks: Semantic Compilation and Latent Memory [5.649244092106778]
  We prove inference-time quantum coordination advantages for specified AI state-tracking tasks.<n>We map a finite one-way, streaming, or adaptive causal task into a semantic AI interface while preserving event order and access to past input.
  arXiv  Detail & Related papers  (2026-08-11T15:32:15Z)
• Sparse Quantum State Preparation with Sublinear T-Count [5.929956715430168]
  We show that any $n$-qubit $s$-sparse state can be prepared up to error $$ using $widetildeO(mins, n3/4sqrts+sqrtslog (1/)+log (1/)$ $T$ gates.<n>For every $0le$ and $2le sle 2n/2$, sparse-state preparation requires $(mins,sqrtns)$ $T
  arXiv  Detail & Related papers  (2026-08-01T03:33:19Z)
• Efficiently Representing Algorithms With Chain-of-Thought Transformers [45.60818649759173]
  Chain-of-thought transformers can simulate Turing machines, and thus perform arbitrary computation.<n>We show that CoT emphcan efficiently simulate any Word RAM algorithm with only a poly-logarithmic overhead in $n$.
  arXiv  Detail & Related papers  (2026-06-18T01:49:29Z)
• Efficient Quantum State Synthesis with One Query [0.0]
  We present a time analogue quantum algorithm making a single query (in superposition) to a classical oracle. We prove that every $n$-qubit state can be constructed to within 0.01 error by an $On/n)$-size circuit over an appropriate finite gate set.
  arXiv  Detail & Related papers  (2023-06-02T17:49:35Z)
• Spacetime-Efficient Low-Depth Quantum State Preparation with Applications [93.56766264306764]
  We show that a novel deterministic method for preparing arbitrary quantum states requires fewer quantum resources than previous methods. We highlight several applications where this ability would be useful, including quantum machine learning, Hamiltonian simulation, and solving linear systems of equations.
  arXiv  Detail & Related papers  (2023-03-03T18:23:20Z)
• Simplifying and Understanding State Space Models with Diagonal Linear RNNs [56.33053691749856]
  This work disposes of the discretization step, and proposes a model based on vanilla Diagonal Linear RNNs. We empirically show that, despite being conceptually much simpler, $mathrmDLR$ is as performant as previously-proposed SSMs. We also characterize the expressivity of SSMs and attention-based models via a suite of $13$ synthetic sequence-to-sequence tasks.
  arXiv  Detail & Related papers  (2022-12-01T18:53:06Z)
• Horizon-Free and Variance-Dependent Reinforcement Learning for Latent Markov Decision Processes [62.90204655228324]
  We study regret minimization for reinforcement learning (RL) in Latent Markov Decision Processes (LMDPs) with context in hindsight. We design a novel model-based algorithmic framework which can be instantiated with both a model-optimistic and a value-optimistic solver.
  arXiv  Detail & Related papers  (2022-10-20T21:32:01Z)
• Near-Optimal Regret Bounds for Multi-batch Reinforcement Learning [54.806166861456035]
  We study the episodic reinforcement learning (RL) problem modeled by finite-horizon Markov Decision Processes (MDPs) with constraint on the number of batches. We design a computational efficient algorithm to achieve near-optimal regret of $tildeO(sqrtSAH3Kln (1/delta))$tildeO(cdot) hides logarithmic terms of $(S,A,H,K)$ in $K$ episodes. Our technical contribution are two-fold: 1) a near-optimal design scheme to explore
  arXiv  Detail & Related papers  (2022-10-15T09:22:22Z)
• An Optimal Separation of Randomized and Quantum Query Complexity [67.19751155411075]
  We prove that for every decision tree, the absolute values of the Fourier coefficients of a given order $ellsqrtbinomdell (1+log n)ell-1,$ sum to at most $cellsqrtbinomdell (1+log n)ell-1,$ where $n$ is the number of variables, $d$ is the tree depth, and $c>0$ is an absolute constant.
  arXiv  Detail & Related papers  (2020-08-24T06:50:57Z)

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.