中 EN
Sign In Register

Academic Intelligence · Curated Daily

Explore the Frontier of Global Academia

AcademicHub aggregates real-time literature from top journals and preprint platforms. Build your personal research radar and let large language models compile cross-disciplinary analysis briefings automatically.

Sign In Register
01.
arXiv (quant-ph) 2026-06-16

Theory of the correlated quantum Zeno effect in a monitored qubit dimer

Authors:
Severino Zeni↗Gobinda Chakraborty↗Alessandro Romito↗Alberto Biella↗

arXiv:2503.22846v2 Announce Type: replace Abstract: We theoretically investigate the stochastic dynamics of two qubits subject to one- and two-site correlated continuous weak measurements. When measurements dominate over the local unitary evolution, the system's dynamics is constrained and part of the physical Hilbert space becomes inaccessible: a typical signature of the Quantum Zeno (QZ) effect. In this work, we show how the competition between these two measurement processes give rise to two distinct QZ regimes, we dubbed standard and correlated, characterised by a different topology of the allowed region of the physical Hilbert space being a simply and non-simply connected domain, respectively. We develop a theory based on a stochastic Gutzwiller ansatz for the wavefunction that is able to capture the structure of the phase diagram. Finally we show how the two QZ regimes are intimately connected to the topology of the flow of the underlying non-Hermitian Hamiltonian governing the no-click evolution.

Read & Discuss → View Source →
03.
arXiv (CS.LG) 2026-06-16

AREAL-DTA: Dynamic Tree Attention for Efficient Reinforcement Learning of Large Language Models

Authors:
Jiarui Zhang↗Yuchen Yang↗Ran Yan↗Zhiyu Mei↗Liyuan Zhang↗Daifeng Li↗Wei Fu↗Jiaxuan Gao↗Shusheng Xu↗Yi Wu↗Binhang Yuan↗

arXiv:2602.00482v2 Announce Type: replace Abstract: Reinforcement learning (RL)-based post-training for large language models (LLMs) is computationally expensive, as it generates many rollout sequences that frequently share long token prefixes. Existing RL frameworks usually process these sequences independently during policy training, i.e., repeatedly recomputing identical prefixes in both the forward and backward passes of policy gradient computation, leading to substantial inefficiencies in computation resources and memory usage. Although prefix sharing naturally induces a tree structure over rollouts, packed tree-mask approaches scale poorly in RL settings. In this paper, we introduce AReaL-DTA, which efficiently exploits prefix sharing in RL training. AReaL-DTA employs a depth-first search (DFS)-based execution strategy that dynamically traverses the rollout prefix tree during both forward and backward computation, materializing only a single root-to-leaf path at a time. To further improve scalability, AReaL-DTA incorporates a load-balanced distributed batching mechanism that dynamically constructs and processes prefix trees across multiple GPUs. On $\tau^2$-bench, AReaL-DTA improves training throughput by up to $8.31\times$ over dense training and up to $1.70\times$ over sparse training. Our code is available at https://github.com/areal-project/AReaL/tree/feat/dta.

Read & Discuss → View Source →
04.
arXiv (CS.AI) 2026-06-15

When the Tool Decides: LLM Agents Defer Blindly to Graph Neural Network Tools, and Stronger Backbones Defer More

Authors:
Zhongyuan Wang↗Pratyusha Vemuri↗

arXiv:2606.14476v1 Announce Type: new Abstract: A growing line of work equips large language model (LLM) agents with graph neural networks (GNNs) as callable tools, assuming the agent exercises judgment over when and how much to rely on such a tool. We test this directly. We expose a frozen GNN to a ReAct-style LLM agent as an explicit tool and measure, on node classification over a text-attributed graph (ogbn-arxiv, replicated on WikiCS), whether the agent uses the tool or merely obeys it. We find the agent does not exercise judgment: its predictions agree with the raw GNN's 97.6-99.2% of the time (5 seeds), collapsing into a GNN parrot that adopts the tool's output wholesale and bypasses its own reasoning. Sweeping backbone capability (Qwen2.5 0.5B-7B), the deference is not a weak-model artifact: among models able to invoke the tool, agreement rises with capability (0.60 to 0.98 from 1.5B to 7B). Crucially, the cost of deference does not shrink as capability grows and grows where alternatives emerge: a per-node oracle over the available actions beats the parrot by 0.09-0.18 at 3B and 0.12-0.22 at 7B, roughly doubling at high homophily, because the parrot is pinned to the frozen GNN while the agent's alternatives improve; at 7B a simple neighbour-label tool overtakes the GNN at high homophily (0.81 vs 0.71) yet the agent still defers. A simple selective-invocation gate recovers about half of that high-homophily gap (0.71 to 0.83) but yields no net global gain, and held-out estimates bound the best achievable gate over standard test-time features to at most a third of the oracle headroom: reliable selective invocation looks limited by available information, not merely router design. Our results are a cautionary measurement: evaluations of agent+tool systems cannot assume the agent adds judgment on top of the tool, and selective invocation must be designed in rather than expected to emerge from scale.

Read & Discuss → View Source →
05.
arXiv (CS.CL) 2026-06-18

Simulating Hate Speech Cascades with Multi-LLM Agents: Empirical Grounding, Modeling Fidelity, and Intervention Strategies

Authors:
Fan Huang↗

Faithful modeling of hateful content propagation on online platforms remains an open problem for moderation research. Classical cascade models that do not explicitly represent the profile, community, and content factors associated with hateful-content propagation may yield moderation strategies that behave less effectively when deployed in real-world scenarios. Multi-agent large language model (LLM) systems can, in principle, make each reshare decision depend on the user's profile, the surrounding community, and the post's content, but it remains unclear whether this added flexibility actually reproduces real hateful cascades more faithfully than classical baselines. We study three hateful Bluesky cascades and a size-matched benign control. In the empirical Bluesky data, we found that: 97.4–99.7\% of reposters take a hostile stance; toxicity-engagement homophily is higher on the diffusion tree than on the follower graph for hateful cascades; topology is star-like for the hateful cascades (most reposts come directly from the root) versus tree-like for the benign cascade (reposts propagate through multi-hop chains). In simulation, a multi-LLM-agent simulator reproduces the stance monoculture and the toxicity-delta direction. A structured ablation identifies agent heterogeneity as the leading fidelity factor, and amplifier targeting on dense networks yields 7.5–12.9\% reduction at 5.7\% benign collateral.

Read & Discuss → View Source →
06.
arXiv (CS.LG) 2026-06-18

Learning Augmented Exact Exponential Algorithms

Authors:
Tatiana Belova↗Yuriy Dementiev↗Danil Sagunov↗

arXiv:2606.18807v1 Announce Type: cross Abstract: The field of learning-augmented algorithms has demonstrated that machine-learned predictions can bypass worst-case lower bounds across a wide range of problems. So far, however, the focus has been almost exclusively on polynomial-time algorithms, where predictions improve competitive ratios, approximation guarantees, or running times. In this paper, we raise the question of whether predictions can push the frontier of exact exponential-time algorithms for NP-hard problems. We answer this question affirmatively by proposing a general approach that augments an entire family of state-of-the-art exact algorithms for a variety of subset selection problems. We show that a noisy predictor that is only marginally better than random guessing suffices to provably reduce the search space, and that the resulting runtime speedup scales smoothly with the prediction quality. Importantly, our algorithms require only pairwise independence of predictions or, alternatively, do not require the knowledge of the predictor's accuracy - both strictly weaker and more realistic settings than typically assumed.

Read & Discuss → View Source →
07.
arXiv (CS.CV) 2026-06-24

Agentic Collaborative Cognition for Zero-Shot 3D Understanding

Authors:
Wenxin Wang↗Bo Zhang↗Feng Chen↗Zixuan Wang↗Wen Li↗Changsheng Li↗Yinjie Lei↗

Recent advancements have explored agentic zero-shot 3D understanding by reformulating it as video keyframe understanding with Multimodal Large Language Models (MLLMs). However, existing methods face an intrinsic bottleneck due to the finite observation perspectives inherent in videos and the implicit perception of 3D scenes. In this paper, we propose a collaborative multi-agent framework that assigns a Planning Agent to handle high-level viewpoint planning and supplement novel perspectives, and a Perception Agent to explicitly summarize the 3D scene into a structured holistic cognitive map. Specifically, Planning Agent first analyzes this cognitive map to determine query-relevant viewpoints and supplements missing critical perspectives to ensure comprehensive observation. Subsequently, Perception Agent documents object-level attributes from these views by assigning consistent instance identifiers across viewpoints, thereby integrating fragmented observations into the holistic cognitive map. In parallel, it provides feedback to filter out mismatched candidate objects and guide subsequent viewpoint planning. Through this closed-loop iterative process, two agents collaboratively figure out candidates until Perception Agent determines that sufficient information has been captured to complete the task. Extensive experiments demonstrate that our method achieves state-of-the-art performance on 6 benchmarks, with improvements of 11.1\% Acc@0.5 on ScanRefer, 14.6 BLEU-1 on 3D-assisted dialog, and 2.1 EM on SQA3D.

Read & Discuss → View Source →
08.
arXiv (quant-ph) 2026-06-11

Planted-Solution Pauli Hamiltonians as a Quantum Benchmarking Primitive

Authors:
Amir Kalev↗Itay Hen↗

arXiv:2606.11455v1 Announce Type: new Abstract: We introduce a construction of Pauli Hamiltonians with exactly known ground-state energies, intended as reference instances for ground-state energy estimation algorithms. The construction embeds a planted block-product state as the simultaneous ground state of a sum of frustration-free local clauses on overlapping supports, exposes the resulting model only as a polynomial-size linear combination of Pauli operators, and admits optional Clifford conjugation that preserves the spectrum. The framework subsumes classical planted constraint-satisfaction problems as a diagonal special case, providing a direct embedding channel through which classical hardness properties can be inherited. Open-source software, certification keys, and example instances are made publicly available.

Read & Discuss → View Source →
09.
arXiv (CS.AI) 2026-06-18

Conflict-Aware Retriever Editing for Knowledge Injection Attacks on LLM-Based RAG Systems

Authors:
Xinru Liu↗Xianglong Zhang↗Di Cai↗Zhumin Chen↗Pengfei Hu↗Xin Xin↗

arXiv:2606.18310v1 Announce Type: cross Abstract: Injecting malicious knowledge into retrieval-augmented generation (RAG) systems can manipulate retrieved evidence and mislead downstream generation, posing a serious security threat for AI applications. Existing RAG injection attacks mainly rely on manipulating external knowledge bases, such as crafting malicious corpus. However, the synthetic text crafted by such data-centric methods could be detectable, leading to the failure of attacks. Beyond corpus manipulation, open-source retrievers are increasingly exposing RAG systems to model-centric attacks. In this paper, we propose conflict-aware retriever editing, i.e., CAREATTACK, a model-centric retriever attack framework for malicious knowledge injection in RAG. Specifically, CAREATTACK consists two stages of conflict-aware retriever editing and attack-preserving anchor repair. Conflict-aware retriever editing adapts efficient closed-form parameter editing to the dense retrieval model, promoting malicious knowledge above benign competing passages and resolving potential parameter conflicts through graph-based conflict detection and parameter editing projection. Then, attack-preserving anchor repair performs lightweight calibration on the edited retriever to further eliminate the impact on non-target prompts while preserving the attack effectiveness for target prompts. We instantiate CAREATTACK on Qwen3-Embedding-0.6B and BGE-M3, and conduct evaluation on three benchmark datasets. Experimental results demonstrate our method substantially promote malicious passages into the retrieved knowledge of RAG systems and can perform attacks for batches of target prompts and passages, given the access of retrieval model parameters. Since most RAG systems are built upon open-source retrieval models, this work reveals a practical attack surface in RAG systems. Codes are public accessible at https://anonymous.4open.science/r/CareAttack-3F1C.

Read & Discuss → View Source →
10.
arXiv (CS.LG) 2026-06-17

Finite-Time Queue Peak Laws in Stochastic Networks: Logarithmic Scaling After Geometric Thresholds

Authors:
Hao Liang↗Cheng Tang↗Yunzong Xu↗

arXiv:2606.18218v1 Announce Type: cross Abstract: We study finite-horizon queue peaks in generalized switches, a standard stochastic-network model in which many queues share constrained service resources. Arrivals may be dependent, time-varying, and adapted to the past; the standing load condition is uniform interior slack, meaning the conditional mean arrival vector stays in a fixed contraction of the capacity region. We show that this slack reshapes the finite-time peak law for drift-minimizing scheduling policies such as MaxWeight. The square-root envelope that is sharp without slack persists only up to a geometry-dependent threshold; beyond that threshold, the running maximum grows only logarithmically with the horizon, both with high probability and in expectation. The mechanism is self-normalization: in the current queue direction, the projected fluctuation scale is normalized by the stabilizing drift scale. This removes capacity geometry from the logarithmic coefficient, while geometry remains in the threshold. Matching lower bounds show that both the logarithmic term and a geometric threshold are unavoidable. When finite-time state-space collapse is available, the threshold can be sharpened using local bottleneck geometry. For generalized input-queued switches, we obtain finite-time peak bounds with tight logarithmic coefficients. Simulations illustrate the two-phase envelope, local geometric refinements, and variance-sensitive improvements predicted by the theory.

Read & Discuss → View Source →
11.
arXiv (CS.CL) 2026-06-12

Multi-Bitwidth Quantization for LLMs Using Additive Codebooks

Authors:
Liza Babaoglu↗Shuangyi Chen↗Ashish Khisti↗

As large language models (LLMs) are increasingly deployed across heterogeneous hardware with varying resource constraints, the ability to adaptively manage the trade-off between performance and efficiency without retraining is critical. We propose Drop-by-Drop, a novel multi-bitwidth post-training quantization framework that enables inference-time precision control over LLM weights from a single trained model. Our method is theoretically grounded in information theory and successive refinement. We establish that LLM weights, which commonly follow a Gaussian distribution, can be optimally reconstructed with increasing fidelity as additional bits are incorporated, under a weighted mean squared error distortion motivated by LLM loss functions. To realize this in practice, Drop-by-Drop incorporates Matryoshka-style supervision into the loss function, exploiting the structure of additive codebooks. Drop-by-Drop produces a single model where ordered subsets of codebooks yield accurate partial reconstructions at each precision level. This approach significantly reduces storage and memory overhead by allowing a single checkpoint to serve multiple bitwidths, while maintaining competitive perplexity and accuracy across major architectures, such as Qwen, LLaMA, Gemma, and Mistral.

Read & Discuss → View Source →
12.
arXiv (quant-ph) 2026-06-24

Low Spatial Cost CCZ Magic State Factory

Authors:
Sungyeon Kook↗Yujin Kang↗Ilkwon Sohn↗Jun Heo↗

arXiv:2606.24170v1 Announce Type: new Abstract: We propose a design framework for reconstructing gate-based magic state distillation protocols as compact joint-measurement architectures implementable with the surface code. The goal is to reduce the surface-code resource cost of a magic state factory while preserving the logical function and error-detection structure of the distillation protocol. We construct a reduced architecture for implementing an eight-to-three CCZ distillation protocol using smaller surface-code patches. The proposed factory preserves the single-fault-detection property and the leading-order error suppression of the protocol, while producing CCZ magic states with lower spatial cost than the design of Gidney and Fowler. The proposed design perspective can also be applied to T-state factories and other multiqubit non-Clifford resource-state factories. Our approach provides a framework for extending the design space of surface-code magic state factories beyond a single CCZ layout optimization.

Read & Discuss → View Source →
13.
arXiv (CS.LG) 2026-06-24

Asymptotic Signal Subspace Recovery in Softmax Attention Models

Authors:
Lan V. Truong↗

arXiv:2606.22406v2 Announce Type: replace Abstract: Attention mechanisms have demonstrated remarkable empirical success in identifying relevant information from large collections of tokens, yet the theoretical principles underlying this behavior remain poorly understood. We study a stylized softmax-attention model in which a query vector is learned by stochastic gradient ascent from a collection of informative and nuisance tokens. Exploiting the symmetry of the model, we derive a population objective and characterize the limiting ordinary differential equation governing the learning dynamics. Using tools from stochastic approximation and dynamical systems theory, we establish a rigorous connection between the stochastic learning algorithm and its deterministic limit. Our main result shows that, under suitable high-dimensional scaling assumptions and standard step-size conditions, the learned query converges almost surely to the one-dimensional signal subspace spanned by the latent informative direction. Equivalently, the query asymptotically recovers the latent signal up to the intrinsic sign ambiguity. These results provide a rigorous theoretical foundation for understanding attention mechanisms as signal extraction procedures in high-dimensional noisy environments and offer a dynamical-systems perspective on how attention discovers relevant information in the presence of substantial noise.

Read & Discuss → View Source →
14.
PLOS Computational Biology 2026-06-01

Histology-informed spatial domain identification through multi-view graph convolutional networks

Authors:
Huihui Zhang↗

by Huihui Zhang, Jiaxing Chang, Zirong Li, Yue Sun, Pinli Hu, Haoxiu Wang, Hang Yang, Yonglin Ren, Xingtan Zhang, Zehua Chen, Kok Wai Wong, Haojing Shao Identifying spatial domains is crucial in spatial transcriptomics, yet effectively integrating gene expression, spatial location, and histology remains challenging. We present STESH, a Spatial Transcriptomics clustering method that combines Expression, Spatial information and Histology. STESH extracts histological features using a convolutional neural network and generates expression, histology, spatial, and collaborative convolution modules for a multi-view graph convolutional network with a decoder and attention mechanism. We evaluated STESH on multiple tissue types and technology platforms. STESH consistently outperformed ten state-of-the-art methods, achieving superior clustering accuracy with the highest scores in adjusted Rand index, normalized mutual information, and Fowlkes-Mallows index.

Read & Discuss → View Source →
15.
arXiv (CS.AI) 2026-06-18

A Convex Route to Thermoelasticity: Learning Internal Energy and Dissipation

Authors:
Hagen Holthusen↗Paul Steinmann↗Ellen Kuhl↗

arXiv:2603.28707v3 Announce Type: replace-cross Abstract: We present a physics-based neural network framework for the discovery of constitutive models in fully coupled thermomechanics. In contrast to classical formulations based on the Helmholtz energy, we adopt the internal energy and a dissipation potential as primary constitutive functions, expressed in terms of deformation and entropy. This choice avoids the need to enforce mixed convexity–concavity conditions and facilitates a consistent incorporation of thermodynamic principles. In this contribution, we focus on materials without preferred directions or internal variables. While the formulation is posed in terms of entropy, the temperature is treated as the independent observable, and the entropy is inferred internally through the constitutive relation, enabling thermodynamically consistent modeling without requiring entropy data. Thermodynamic admissibility of the networks is guaranteed by construction. The internal energy and dissipation potential are represented by input convex neural networks, ensuring convexity and compliance with the second law. Objectivity, material symmetry, and normalization are embedded directly into the architecture through invariant-based representations and zero-anchored formulations. We demonstrate the performance of the proposed framework on synthetic and experimental datasets, including purely thermal problems and fully coupled thermomechanical responses of soft tissues and filled rubbers. The results show that the learned models accurately capture the underlying constitutive behavior. All code, data, and trained models are made publicly available via https://doi.org/10.5281/zenodo.19248596.

Read & Discuss → View Source →
16.
arXiv (CS.CL) 2026-06-12

Reward Modeling for Multi-Agent Orchestration

Authors:
King Yeung Tsang↗Zihao Zhao↗Vishal Venkataramani↗Haizhou Shi↗Zixuan Ke↗Semih Yavuz↗Shafiq Joty↗Hao Wang↗

Multi-Agent Systems (MAS) built on Large Language Models (LLMs) require effective orchestration to coordinate specialized agents, yet training such orchestrators is hindered by limited supervision and high computational cost. We propose Orchestration Reward Modeling (OrchRM), a self-supervised framework for evaluating orchestration quality without human annotations. OrchRM leverages intermediate artifacts from multi-agent executions to construct win-lose pairs for Bradley-Terry reward model training. Unlike existing MAS test-time scaling and orchestrator training frameworks that rely on costly sub-agent rollouts, OrchRM operates directly at the orchestration level, enabling efficient and high-performing reward-guided orchestrator training and MAS test-time scaling. OrchRM improves training efficiency by up to 10x in token usage while improving MAS test-time scaling performance by up to 8% in accuracy. These gains consistently transfer across multiple domains, including mathematical reasoning, web-based question answering, and multi-hop reasoning, demonstrating orchestration-level reward modeling as a scalable direction for robust multi-agent orchestration. Code will be available at https://github.com/Wang-ML-Lab/OrchRM.

Read & Discuss → View Source →
17.
bioRxiv (Bioinfo) 2026-06-14

Transposable elements as evolutionary substrates of proteindisorder in the human proteome

Authors:
Mac Donagh↗Vergesio↗Aguilar↗Nores↗Lagares↗Fornasari↗M. S↗Parisi↗

Intrinsically disordered regions (IDRs) are central contributors to protein function, evolution and human disease, yet the evolutionary routes that seed new disordered segments within pre-existing proteins are still poorly understood. Sequence insertions provide a powerful mechanism for disorder expansion, but the genomic donors of inserted IDR and its long-term conformational fate remain largely unknown. Transposable elements (TEs), abundant mobile genetic elements with distinctive compositional biases, represent compelling candidates for generating disorder within proteins. Here, we systematically mapped TE-derived segments across human proteins and isoforms, and we found that these insertions are strongly enriched in intrinsic disorder. The structural consequences of their insertion are shaped by TE class and family, reflecting the sequence biases of the elements from which they originate. Recent, Primate specific insertions preferentially generate disordered segments, whereas older insertions more frequently occupy ordered structural contexts, revealing an age-dependent transition in the conformational state of TE-derived sequences. TE-containing isoforms are expressed at lower levels than TE-free isoforms, particularly when insertions are young and disorder-rich, suggesting that intrinsic disorder may constrain the cellular tolerance of newly exonized sequences. These findings identify TEs as a major evolutionary mechanism linking genome mobility to the emergence of new disordered conformational ensembles in the human proteome.

Read & Discuss → View Source →
18.
arXiv (CS.AI) 2026-06-15

Metabolic cost of information processing in Poisson variational autoencoders

Authors:
Hadi Vafaii↗Jacob L. Yates↗

arXiv:2602.13421v2 Announce Type: replace-cross Abstract: Computation in biological systems is fundamentally energy-constrained, yet standard theories of computation treat energy as freely available. Here, we argue that variational free energy minimization under a Poisson assumption offers a principled path toward an energy-aware theory of computation. Our key observation is that the Kullback-Leibler (KL) divergence term in the Poisson free energy objective becomes proportional to the prior firing rates of model neurons, yielding an emergent metabolic cost term that penalizes high baseline activity. This structure couples an abstract information-theoretic quantity – the *coding rate* – to a concrete biophysical variable – the *firing rate* – which enables a trade-off between coding fidelity and energy expenditure. Such a coupling arises naturally in the Poisson variational autoencoder (P-VAE) – a brain-inspired generative model that encodes inputs as discrete spike counts and recovers a spiking form of *sparse coding* as a special case – but is absent from standard Gaussian VAEs. To demonstrate that this metabolic cost structure is unique to the Poisson formulation, we compare the P-VAE against Grelu-VAE, a Gaussian VAE with ReLU rectification applied to latent samples, which controls for the non-negativity constraint. Across a systematic sweep of the KL term weighting coefficient $\beta$ and latent dimensionality, we find that increasing $\beta$ monotonically increases sparsity and reduces average spiking activity in the P-VAE. In contrast, Grelu-VAE representations remain unchanged, confirming that the effect is specific to Poisson statistics rather than a byproduct of non-negative representations. These results establish Poisson variational inference as a promising foundation for a resource-constrained theory of computation.

Read & Discuss → View Source →
19.
arXiv (CS.AI) 2026-06-15

Distributional Biases in Post-Training: A Markovian Analysis of Reasoning Trajectories

Authors:
Dake Bu↗Wei Huang↗Andi Han↗Atsushi Nitanda↗Bo Xue↗Qingfu Zhang↗Hau-San Wong↗Taiji Suzuki↗

arXiv:2511.07368v3 Announce Type: replace-cross Abstract: Foundation models exhibit broad knowledge but limited task-specific reasoning, motivating post-training strategies such as RL with verifiable rewards (RLVR) and test-time scaling (TTS). While recent work highlights the role of exploration in improving pass@K, empirical evidence points to a paradox: RLVR and ORM/PRM typically reinforce existing paths rather than expanding the reasoning scope, raising the question of why exploration helps if no new patterns emerge. To reconcile this paradox, we adopt the perspective of Kim et al. (2025), viewing easy (e.g., simplifying a fraction) versus hard (e.g., discovering the some symmetry) reasoning steps as low versus high probability Markov transitions. In this tractable model, pretraining corresponds to tree-graph discovering, while post-training corresponds to CoT reweighting. We provably show that, both RLVR and ORM/PRM would favor heavily to several high-probability paths, and thereby forget rare-but-crucial CoTs. Building on this, we further prove that exploration strategies such as rejecting easy instances and KL regularization help preserve rare CoTs. Empirical simulations corroborate our theoretical results.

Read & Discuss → View Source →
20.
arXiv (CS.AI) 2026-06-24

Can Scale Save Us From Plasticity Loss in Large Language Models?

Authors:
J. Fernando Hernandez-Garcia↗Tom\'as Figliolia↗Beren Millidge↗

arXiv:2606.24752v1 Announce Type: new Abstract: The loss of plasticity - the ability of a network to learn new information after having already learned older information - is a fundamental challenge in creating artificial neural networks capable of continual learning. Although this phenomenon has been known for decades, it has mostly been studied in older, relatively small architectures and rarely in natural-language domains. To determine whether loss of plasticity remains a problem in the modern transformer-based LLM paradigm, we study plasticity loss in GPT-style Transformer models trained on a multilingual continual learning problem. Consistent with prior work, we find evidence of plasticity loss across models ranging from 5M to 314M non-embedding parameters, as measured by deterioration on a held-out Vietnamese probing task. We further find that the onset of plasticity loss follows a predictable scaling law, growing sublinearly with model size. These results suggest that larger models may delay the measurable effects of plasticity loss, but that increasing parameter count alone is likely to be insufficient to completely prevent it. We also find evidence of plasticity loss under stationary multilingual training, challenging the view that the phenomenon is exclusive to continual learning with abrupt task changes. Overall, our results suggest that even large Transformer language models trained on natural-language will eventually lose the ability to efficiently adapt to new data after sufficiently long training, in both continual and stationary settings.

Read & Discuss → View Source →
21.
arXiv (CS.AI) 2026-06-12

Select and Improve: Understanding the Mechanics of Post-Training for Reasoning

Authors:
Akshay Krishnamurthy↗Audrey Huang↗Nived Rajaraman↗

arXiv:2606.13125v1 Announce Type: cross Abstract: Reinforcement learning has rapidly emerged as a key component in the training of reasoning and coding models, yet it remains poorly understood from a mechanistic perspective. We study how and through what underlying processes capabilities are acquired or enhanced via reinforcement learning post-training. Our analysis, based on controlled math reasoning experiments with Qwen-2.5-1.5B, reveals two core mechanisms: strategy selection and strategy improvement. Our results highlight the role of SFT data and reinforcement learning data in activating these mechanisms, in particular showing how supervising the model on diverse reasoning strategies can enable strategy selection and how increasing difficulty in reinforcement learning data can enable strategy improvement. Taken together, our results provide mechanistic insight into RL training and suggest practical interventions to continue scaling reasoning capabilities.

Read & Discuss → View Source →
22.
arXiv (CS.AI) 2026-06-11

Improving Generalization and Data Efficiency with Diffusion in Offline Multi-agent RL

Authors:
Zhuoran Li↗Ling Pan↗Jiatai Huang↗Longbo Huang↗

arXiv:2307.01472v2 Announce Type: replace Abstract: We present a novel Diffusion Offline Multi-agent Model (DOM2) for offline Multi-Agent Reinforcement Learning (MARL). Different from existing algorithms that rely mainly on conservatism in policy design, DOM2 enhances policy expressiveness and diversity based on diffusion model. Specifically, we incorporate a diffusion model into the policy network and propose a trajectory-based data-reweighting scheme in training. These key ingredients significantly improve algorithm robustness against environment changes and achieve significant improvements in performance, generalization and data-efficiency. Our extensive experimental results demonstrate that DOM2 outperforms existing state-of-the-art methods in all multi-agent particle and multi-agent MuJoCo environments, and generalizes significantly better to shifted environments {(in $28$ out of $30$ settings evaluated)} thanks to its high expressiveness and diversity. Moreover, DOM2 is ultra data efficient and requires no more than $5\%$ data for achieving the same performance compared to existing algorithms (a $20\times$ improvement in data efficiency).

Read & Discuss → View Source →
23.
arXiv (CS.LG) 2026-06-16

Floating-Point Networks with Automatic Differentiation Can Represent Almost All Floating-Point Functions and Their Gradients

Authors:
Sejun Park↗Yeachan Park↗Geonho Hwang↗

arXiv:2605.01702v2 Announce Type: replace Abstract: Theoretical studies show that for any differentiable function on a compact domain, there exists a neural network that approximates both the function values and gradients. However, such a result cannot be used in practice since it assumes real parameters and exact internal operations. In contrast, real implementations only use a finite subset of reals and machine operations with round-off errors. In this work, we investigate whether a similar result holds for neural networks under floating-point arithmetic, when the gradient with respect to the input is computed by the automatic differentiation algorithm $D^\mathtt{AD}$. We first show that given a floating-point function $\phi$ (e.g., a loss function), arbitrary function values and gradients can be represented by a floating-point network $f$ and $D^\mathtt{AD}(\phi\circ f)$, respectively. We further extend this result: given $\phi_1,\dots,\phi_n$, $D^\mathtt{AD}(\phi_i\circ f)$ can simultaneously represent arbitrary gradients while $f$ represents the target values, under mild conditions. Our results hold for practical activation functions, e.g., $\mathrm{ReLU}$, $\mathrm{ELU}$, $\mathrm{GeLU}$, $\mathrm{Swish}$, $\mathrm{Sigmoid}$, and $\mathrm{tanh}$.

Read & Discuss → View Source →
24.
arXiv (CS.AI) 2026-06-12

TimeROME-DLM: Temporal Causal Tracing and Low-Rank Inference-Time Knowledge Editing for Masked Diffusion Language Models

Authors:
Zhengtao Yao↗Liuyang Song↗Hongbo Zhang↗Chenhao Wei↗Haoyan Xu↗Guang Yang↗Siheng Wang↗

arXiv:2606.12841v1 Announce Type: cross Abstract: Masked diffusion language models (MDLMs) such as LLaDA now rival autoregressive (AR) LLMs, but every existing knowledge-editing and unlearning method (ROME, MEMIT, etc.) targets AR transformers and either makes assumptions that fail under iterative denoising, or requires gradient updates whose backward-pass activations cost tens of GB of extra VRAM and which collapse MDLMs at standard learning rates. We introduce TimeROME-DLM, the first training-free, gradient-free, inference-time knowledge-editing framework for MDLMs. It couples two components: a Temporal Indirect Effect (TIE) causal-tracing protocol that identifies, for each fact, the coordinate whose intervention most strongly drives the object prediction at later denoising steps; and a closed-form, low-rank residual edit memory that aggregates subject keys and target deltas across all forget facts and applies a single ridge-regularised update at that coordinate at every diffusion forward, with sparsification to limit utility spillover. Backbone weights stay frozen; only three hyperparameters (alpha, lambda, q) are tuned on a small validation split. On TOFU forget01 with TOFU-finetuned LLaDA-8B-Base, TimeROME-DLM cuts forget-set log-probability by roughly 83 nats. The same configuration transfers to LLaDA-8B-Instruct, Dream-7B, MMaDA-8B, DiffuLLaMA-7B, and LLaDA-MoE-1.4B. It keeps retain-set log-probability nearly flat (within ~1 nat at the utility-safe operating point) across 50 sequentially inserted facts, delivers a four- to fourteen-fold wall-clock speedup with zero additional VRAM over the strongest converged training-time baseline, and scales sub-linearly to 400 facts. TimeROME-DLM closes the locate-then-edit gap between AR LLMs and MDLMs at a fraction of the computational cost.

Read & Discuss → View Source →
25.
PLOS Computational Biology 2026-06-04

Cell differentiation can underpin the reproducibility of morphogenesis

Authors:
Dominic K. Devlin↗

by Dominic K. Devlin, Austen R. D. Ganley, Nobuto Takeuchi Morphogenesis of complex body shapes is reproducible despite the noise inherent in the underlying morphogenetic processes. However, how these morphogenetic processes work together to achieve this reproducibility remains unclear. Here, we ask how this reproducibility is achieved by evolving complex morphologies in a multi-scale, computational model. Each morphology consists of a population of cells on a two-dimensional grid using the Cellular Potts Model framework. Each cell contains a genome that encodes a gene regulatory network, morphogens for cell-cell signalling, and proteins that determine cell behaviours. By repeatedly simulating our model with different initial conditions under selection for shape complexity, we obtained a “zoo” of evolved morphologies. We find that these evolved, complex morphologies are reproducible in a sizeable fraction of simulations, despite no direct selection for reproducibility. We show that high reproducibility is caused by spatially segregating moving cells that “shape” morphologies from stationary cells that “maintain” morphologies during morphogenesis. Strikingly, most highly reproducible morphologies also evolved cell differentiation, where proliferative, moving progenitor cells irreversibly differentiate into non-dividing, stationary differentiated cells at tissue boundaries. These results suggest that cell differentiation observed in natural development plays a fundamental role in morphogenesis in addition to the production of specialised cell types. This previously unrecognised role of cell differentiation has major implications for our understanding of how morphologies are generated and regenerated.

Read & Discuss → View Source →