Tootfinder

DeSamba: Decoupled Spectral Adaptive Framework for 3D Multi-Sequence MRI Lesion Classification
Dezhen Wang, Sheng Miao, Rongxin Chai, Jiufa Cui
https://arxiv.org/abs/2507.15487

DeSamba: Decoupled Spectral Adaptive Framework for 3D Multi-Sequence MRI Lesion Classification
Magnetic Resonance Imaging (MRI) sequences provide rich spatial and frequency domain information, which is crucial for accurate lesion classification in medical imaging. However, effectively integrating multi-sequence MRI data for robust 3D lesion classification remains a challenge. In this paper, we propose DeSamba (Decoupled Spectral Adaptive Network and Mamba-Based Model), a novel framework designed to extract decoupled representations and adaptively fuse spatial and spectral features for le…

Mamba-2 audio captioning: design space exploration and analysis
Taehan Lee, Jaehan Jung, Hyukjun Lee
https://arxiv.org/abs/2509.15680 https://arxiv.org/pdf…

Mamba-2 audio captioning: design space exploration and analysis
We present an audio captioning model built on the Mamba-2 large language model backbone, which is a state-of-the-art (SOTA) state-space model (SSM). We systematically explore the design space: LLM sizes, LoRA ranks, and connector designs leveraging Mamba-2's linear-time complexity with respect to sequence length. Across benchmarks, our models achieve strong captioning performance compared with larger language models trained on the same dataset, despite using fewer parameters. For the first time…

Dissociating model architectures from inference computations
Noor Sajid, Johan Medrano
https://arxiv.org/abs/2507.15776 https://arxiv…

Dissociating model architectures from inference computations
Parr et al., 2025 examines how auto-regressive and deep temporal models differ in their treatment of non-Markovian sequence modelling. Building on this, we highlight the need for dissociating model architectures, i.e., how the predictive distribution factorises, from the computations invoked at inference. We demonstrate that deep temporal computations are mimicked by autoregressive models by structuring context access during iterative inference. Using a transformer trained on next-token predict…

Control-Based Online Distributed Optimization
Wouter J. A. van Weerelt, Nicola Bastianello
https://arxiv.org/abs/2508.15498 https://arxiv.org/pdf/2508.1549…

Control-Based Online Distributed Optimization
In this paper we design a novel class of online distributed optimization algorithms leveraging control theoretical techniques. We start by focusing on quadratic costs, and assuming to know an internal model of their variation. In this set-up, we formulate the algorithm design as a robust control problem, showing that it yields a fully distributed algorithm. We also provide a distributed routine to acquire the internal model. We show that the algorithm converges exactly to the sequence of optima…

Universal Dancing by Luminous Robots under Sequential Schedulers
Caterina Feletti, Paola Flocchini, Debasish Pattanayak, Giuseppe Prencipe, Nicola Santoro
https://arxiv.org/abs/2508.15484

Universal Dancing by Luminous Robots under Sequential Schedulers
The Dancing problem requires a swarm of $n$ autonomous mobile robots to form a sequence of patterns, aka perform a choreography. Existing work has proven that some crucial restrictions on choreographies and initial configurations (e.g., on repetitions of patterns, periodicity, symmetries, contractions/expansions) must hold so that the Dancing problem can be solved under certain robot models. Here, we prove that these necessary constraints can be dropped by considering the LUMI model (i.e., wher…

DGenCTR: Towards a Universal Generative Paradigm for Click-Through Rate Prediction via Discrete Diffusion
Moyu Zhang, Yun Chen, Yujun Jin, Jinxin Hu, Yu Zhang
https://arxiv.org/abs/2508.14500

DGenCTR: Towards a Universal Generative Paradigm for Click-Through Rate Prediction via Discrete Diffusion
Recent advances in generative models have inspired the field of recommender systems to explore generative approaches, but most existing research focuses on sequence generation, a paradigm ill-suited for click-through rate (CTR) prediction. CTR models critically depend on a large number of cross-features between the target item and the user to estimate the probability of clicking on the item, and discarding these cross-features will significantly impair model performance. Therefore, to harness t…

Streaming Sequence-to-Sequence Learning with Delayed Streams Modeling
Neil Zeghidour, Eugene Kharitonov, Manu Orsini, V\'aclav Volhejn, Gabriel de Marmiesse, Edouard Grave, Patrick P\'erez, Laurent Mazar\'e, Alexandre D\'efossez
https://arxiv.org/abs/2509.08753

Streaming Sequence-to-Sequence Learning with Delayed Streams Modeling
We introduce Delayed Streams Modeling (DSM), a flexible formulation for streaming, multimodal sequence-to-sequence learning. Sequence-to-sequence generation is often cast in an offline manner, where the model consumes the complete input sequence before generating the first output timestep. Alternatively, streaming sequence-to-sequence rely on learning a policy for choosing when to advance on the input stream, or write to the output stream. DSM instead models already time-aligned streams with a …

Human Feedback Driven Dynamic Speech Emotion Recognition
Ilya Fedorov, Dmitry Korobchenko
https://arxiv.org/abs/2508.14920 https://arxiv.org/pdf/2508.14920…

Human Feedback Driven Dynamic Speech Emotion Recognition
This work proposes to explore a new area of dynamic speech emotion recognition. Unlike traditional methods, we assume that each audio track is associated with a sequence of emotions active at different moments in time. The study particularly focuses on the animation of emotional 3D avatars. We propose a multi-stage method that includes the training of a classical speech emotion recognition model, synthetic generation of emotional sequences, and further model improvement based on human feedback.…

Tactile stimulations reduce or promote the segregation of auditory streams: psychophysics and modeling https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1012701 "cross-modal synchronization, with carefully timed tactile cues, coul…

Tactile stimulations reduce or promote the segregation of auditory streams: psychophysics and modeling
Auditory stream segregation plays a crucial role in understanding the auditory scene. This study investigates the role of tactile stimulation in auditory stream segregation through psychophysics experiments and a computational model of audio-tactile interactions. We examine how tactile pulses, synchronized with one group of tones (high- or low-frequency tones) in a sequence of interleaved high- and low-frequency tones (ABA- triplets), influence the likelihood of perceiving integrated or segrega…

The BRS Cohomology of the Wess Zumino Chiral Scalar supersymmetric model with exotic pairs and exotic triplets (E2)
John A. Dixon
https://arxiv.org/abs/2507.14174 https://

The BRS Cohomology of the Wess Zumino Chiral Scalar supersymmetric model with exotic pairs and exotic triplets (E2)
Using the spectral sequence method, this paper advances some of the construction of the BRS cohomology of the Wess Zumino supersymmetric action. An important missing part was the inclusion of the sources for the variations of the fields. In this paper, these sources are called pseudofields. Since the most interesting part of the result contains unsaturated spinor indices, we include a constant spinor to saturate those indices. At dimension zero, this gives rise to a new set of invariants and a …

SeqVLA: Sequential Task Execution for Long-Horizon Manipulation with Completion-Aware Vision-Language-Action Model
Ran Yang, Zijian An, Lifeng ZHou, Yiming Feng
https://arxiv.org/abs/2509.14138

SeqVLA: Sequential Task Execution for Long-Horizon Manipulation with Completion-Aware Vision-Language-Action Model
Long-horizon robotic manipulation tasks require executing multiple interdependent subtasks in strict sequence, where errors in detecting subtask completion can cascade into downstream failures. Existing Vision-Language-Action (VLA) models such as $π_0$ excel at continuous low-level control but lack an internal signal for identifying when a subtask has finished, making them brittle in sequential settings. We propose SeqVLA, a completion-aware extension of $π_0$ that augments the base architect…

Nature of the Topological Transition of the Kitaev Model in [111] Magnetic Field
S. Thiagarajan, C. Watson, T. Yzeiri, H. Hu, B. Uchoa, F. Kr\"uger
https://arxiv.org/abs/2509.13057

Nature of the Topological Transition of the Kitaev Model in [111] Magnetic Field
We investigate the nature of the topological phase transition of the antiferromagnetic Kitaev model on the honeycomb lattice in the presence of a magnetic field along the [111] direction. The field opens a topological gap in the Majorana fermion spectrum and leads to a sequence of topological phase transitions before the field polarised state is reached. At mean field level the gap first closes at the three $M$ points in the Brillouin zone, where the Majorana fermions form Dirac cones, resultin…

When marine radar target detection meets pretrained large language models
Qiying Hu, Linping Zhang, Xueqian Wang, Gang Li, Yu Liu, Xiao-Ping Zhang
https://arxiv.org/abs/2509.12110

When marine radar target detection meets pretrained large language models
Deep learning (DL) methods are widely used to extract high-dimensional patterns from the sequence features of radar echo signals. However, conventional DL algorithms face challenges such as redundant feature segments, and constraints from restricted model sizes. To address these issues, we propose a framework that integrates feature preprocessing with large language models (LLMs). Our preprocessing module tokenizes radar sequence features, applies a patch selection algorithm to filter out uninf…

OnlineHOI: Towards Online Human-Object Interaction Generation and Perception
Yihong Ji, Yunze Liu, Yiyao Zhuo, Weijiang Yu, Fei Ma, Joshua Huang, Fei Yu
https://arxiv.org/abs/2509.12250

OnlineHOI: Towards Online Human-Object Interaction Generation and Perception
The perception and generation of Human-Object Interaction (HOI) are crucial for fields such as robotics, AR/VR, and human behavior understanding. However, current approaches model this task in an offline setting, where information at each time step can be drawn from the entire interaction sequence. In contrast, in real-world scenarios, the information available at each time step comes only from the current moment and historical data, i.e., an online setting. We find that offline methods perform…

Can any model be fabricated? Inverse operation based planning for hybrid additive-subtractive manufacturing
Yongxue Chen, Tao Liu, Yuming Huang, Weiming Wang, Tianyu Zhang, Kun Qian, Zikang Shi, Charlie C. L. Wang
https://arxiv.org/abs/2509.10599

Can any model be fabricated? Inverse operation based planning for hybrid additive-subtractive manufacturing
This paper presents a method for computing interleaved additive and subtractive manufacturing operations to fabricate models of arbitrary shapes. We solve the manufacturing planning problem by searching a sequence of inverse operations that progressively transform a target model into a null shape. Each inverse operation corresponds to either an additive or a subtractive step, ensuring both manufacturability and structural stability of intermediate shapes throughout the process. We theoretically…

Representing Speech Through Autoregressive Prediction of Cochlear Tokens
Greta Tuckute, Klemen Kotar, Evelina Fedorenko, Daniel L. K. Yamins
https://arxiv.org/abs/2508.11598 htt…

Representing Speech Through Autoregressive Prediction of Cochlear Tokens
We introduce AuriStream, a biologically inspired model for encoding speech via a two-stage framework inspired by the human auditory processing hierarchy. The first stage transforms raw audio into a time-frequency representation based on the human cochlea, from which we extract discrete \textbf{cochlear tokens}. The second stage applies an autoregressive sequence model over the cochlear tokens. AuriStream learns meaningful phoneme and word representations, and state-of-the-art lexical semantics.…

Spatially-Augmented Sequence-to-Sequence Neural Diarization for Meetings
Li Li, Ming Cheng, Hongyu Zhang, Juan Liu, Ming Li
https://arxiv.org/abs/2510.09505 https://

Spatially-Augmented Sequence-to-Sequence Neural Diarization for Meetings
This paper proposes a Spatially-Augmented Sequence-to-Sequence Neural Diarization (SA-S2SND) framework, which integrates direction-of-arrival (DOA) cues estimated by SRP-DNN into the S2SND backbone. A two-stage training strategy is adopted: the model is first trained with single-channel audio and DOA features, and then further optimized with multi-channel inputs under DOA guidance. In addition, a simulated DOA generation scheme is introduced to alleviate dependence on matched multi-channel corp…

Topos Theory for Generative AI and LLMs
Sridhar Mahadevan
https://arxiv.org/abs/2508.08293 https://arxiv.org/pdf/2508.08293

Topos Theory for Generative AI and LLMs
We propose the design of novel categorical generative AI architectures (GAIAs) using topos theory, a type of category that is ``set-like": a topos has all (co)limits, is Cartesian closed, and has a subobject classifier. Previous theoretical results on the Transformer model have shown that it is a universal sequence-to-sequence function approximator, and dense in the space of all continuous functions with compact support on the Euclidean space of embeddings of tokens. Building on this theoretica…

A new deepening of mass-radius empirical relation for main sequence stars
Salvatore Camposeo
https://arxiv.org/abs/2508.08845 https://arxiv.org/pdf/2508.08…

A new deepening of mass-radius empirical relation for main sequence stars
In the following paper I have compared some typical mass-radius relations for main sequence stars by studying their level of agreement with DEBCat [https://doi.org/10.48550/arXiv.1411.1219], which is a recent catalogue by J. Southworth. Models chosen for testing were originally developed using older, smaller datasets than DEBCat. Each model follows a two-piece function structure, where each branch is a monomial power-law. This approach is motivated by theoretical considerations suggesting that …

On Structured State-Space Duality
Jerry Yao-Chieh Hu, Xiwen Zhang, Weimin Wu, Han Liu
https://arxiv.org/abs/2510.04944 https://arxiv.org/pdf/2510.04944

On Structured State-Space Duality
Structured State-Space Duality (SSD) [Dao & Gu, ICML 2024] is an equivalence between a simple Structured State-Space Model (SSM) and a masked attention mechanism. In particular, a state-space model with a scalar-times-identity state matrix is equivalent to a masked self-attention with a $1$-semiseparable causal mask. Consequently, the same sequence transformation (model) has two algorithmic realizations: as a linear-time $O(T)$ recurrence or as a quadratic-time $O(T^2)$ attention. In this note,…

Recovering the Coupled Treatment of Redshift-Space Distortions and the Lightcone Effect after Diffuse Foreground Removal
Jennifer Feron, Emma Chapman
https://arxiv.org/abs/2509.13920

Recovering the Coupled Treatment of Redshift-Space Distortions and the Lightcone Effect after Diffuse Foreground Removal
The 21 cm brightness temperature during the Epoch of Reionisation is widely modelled using semi-numeric simulations, used for their computational speed and flexibility in testing astrophysical and cosmological parameters. However, it is common practice to simulate coeval brightness temperature boxes, and then apply post-processing algorithms that treat the lightcone effect and redshift-space distortions separately, assuming they can be added in sequence. We instead model them together, allowing…

In-Context Learning as Nonparametric Conditional Probability Estimation: Risk Bounds and Optimality
Chenrui Liu, Falong Tan, Chuanlong Xie, Yicheng Zeng, Lixing Zhu
https://arxiv.org/abs/2508.08673

In-Context Learning as Nonparametric Conditional Probability Estimation: Risk Bounds and Optimality
This paper investigates the expected excess risk of In-Context Learning (ICL) for multiclass classification. We model each task as a sequence of labeled prompt samples and a query input, where a pre-trained model estimates the conditional class probabilities of the query. The expected excess risk is defined as the average truncated Kullback-Leibler (KL) divergence between the predicted and ground-truth conditional class distributions, averaged over a specified family of tasks. We establish a ne…

Sequence Reconstruction Problem for Ternary Deletion Channels
Xiang Wang, Han Li, Fang-Wei Fu
https://arxiv.org/abs/2509.24237 https://arxiv.org/pdf/2509.2…

Sequence Reconstruction Problem for Ternary Deletion Channels
The sequence reconstruction problem was proposed by Levenshtein in 2001. In this model, a sequence from a code is transmitted over several channels, and the decoder receives the distinct outputs from each channel. The main problem is to determine the minimum number of channels required to reconstruct the transmitted sequence. In the combinatorial context, the sequence reconstruction problem is equivalent to finding the value of $N_q(n,d,t)$, defined as the size of the largest intersection of tw…

MCeT: Behavioral Model Correctness Evaluation using Large Language Models
Khaled Ahmed, Jialing Song, Boqi Chen, Ou Wei, Bingzhou Zheng
https://arxiv.org/abs/2508.00630 https://…

MCeT: Behavioral Model Correctness Evaluation using Large Language Models
Behavioral model diagrams, e.g., sequence diagrams, are an essential form of documentation that are typically designed by system engineers from requirements documentation, either fully manually or assisted by design tools. With the growing use of Large Language Models (LLM) as AI modeling assistants, more automation will be involved in generating diagrams. This necessitates the advancement of automatic model correctness evaluation tools. Such a tool can be used to evaluate both manually and AI …

Fibonacci Numbers and Model-Complete Axiomatization of Presburger Arithmetic Expanded with a Beatty Sequence
Mohsen Khani, Ali N. Valizadeh, Afshin Zarei
https://arxiv.org/abs/2508.02303

Fibonacci Numbers and Model-Complete Axiomatization of Presburger Arithmetic Expanded with a Beatty Sequence
We introduce a recursive theory that completely axiomatizes the structure $\langle \mathbb{Z},<, +,f,0\rangle$ where $f$ is the function that maps each $x$ to the integer part of $φx $, with $φ$ the golden ratio. We prove that our axiomatization is model-complete in a language expanded with a function which we call the Fibonacci floor function, as defined in the introduction. We also present an improvement of the axiomatization of the structure $\langle \mathbb{Z}, +, f,0\rangle$ already stud…

A new approach to rational stable parametrized homotopy theory
Yves F\'elix, Aniceto Murillo, Alejandro Saiz
https://arxiv.org/abs/2509.11393 https://a…

A new approach to rational stable parametrized homotopy theory
This work develops a comprehensive algebraic model for rational stable parametrized homotopy theory over arbitrary base spaces. Building on the simplicial analogue of the foundational framework of May-Sigurdsson for parametrized spectra, and the homotopy theory of complete differential graded Lie algebras, we construct an explicit sequence of Quillen equivalences that translate the homotopy theory of rational spectra of retractive simplicial sets into the purely algebraic framework of complete …

DeepTrust: Multi-Step Classification through Dissimilar Adversarial Representations for Robust Android Malware Detection
Daniel Pulido-Cort\'azar, Daniel Gibert, Felip Many\`a
https://arxiv.org/abs/2510.12310

DeepTrust: Multi-Step Classification through Dissimilar Adversarial Representations for Robust Android Malware Detection
Over the last decade, machine learning has been extensively applied to identify malicious Android applications. However, such approaches remain vulnerable against adversarial examples, i.e., examples that are subtly manipulated to fool a machine learning model into making incorrect predictions. This research presents DeepTrust, a novel metaheuristic that arranges flexible classifiers, like deep neural networks, into an ordered sequence where the final decision is made by a single internal model…

Electro-viscoelasticity of polymer melts in continuum theory
Zachary Wolfgram, Jeffrey G. Ethier, Matthew Grasinger
https://arxiv.org/abs/2509.13146 https://

Electro-viscoelasticity of polymer melts in continuum theory
Electro-viscoelastic theory for polymer melts has been extensively studied experimentally for the past century, primarily for manufacturing purposes. However, the modeling and theory for this have been minimal, leaving many questions on the mechanisms and behavior of an arbitrary flow scheme. To remedy this, previously solved overdamped Langevin equations for the Doi-Rouse model are modified to include charge and electric field potential forces. The charge sequence on the chain is hypothesized …

TrinityDNA: A Bio-Inspired Foundational Model for Efficient Long-Sequence DNA Modeling
Qirong Yang, Yucheng Guo, Zicheng Liu, Yujie Yang, Qijin Yin, Siyuan Li, Shaomin Ji, Linlin Chao, Xiaoming Zhang, Stan Z. Li
https://arxiv.org/abs/2507.19229

TrinityDNA: A Bio-Inspired Foundational Model for Efficient Long-Sequence DNA Modeling
The modeling of genomic sequences presents unique challenges due to their length and structural complexity. Traditional sequence models struggle to capture long-range dependencies and biological features inherent in DNA. In this work, we propose TrinityDNA, a novel DNA foundational model designed to address these challenges. The model integrates biologically informed components, including Groove Fusion for capturing DNA's structural features and Gated Reverse Complement (GRC) to handle the inhe…

Grid Restoration Under Uncertainty Considering Coupled Transportation-Power Networks
Harshal D. Kaushik, Roshni Anna Jacob, Souma Chowdhury, Jie Zhang
https://arxiv.org/abs/2510.10399

Grid Restoration Under Uncertainty Considering Coupled Transportation-Power Networks
A stochastic mixed-integer programming model is developed to address the power distribution system repair and restoration following failures caused by extreme events such as natural disasters. This model addresses the complex challenge of efficiently assigning and dispatching repair crews to minimize both downtime and the extent of outages, following the crew sequence constraints. By incorporating a realistic transportation network, our model accounts for uncertainties in repair times, repair d…

Service Function Chaining Architecture for Multi-hop Split Inference and Learning
Takanori Hara, Masahiro Sasabe
https://arxiv.org/abs/2509.10001 https://a…

Service Function Chaining Architecture for Multi-hop Split Inference and Learning
Service Function Chaining (SFC) is a networking technique that ensures traffic traverses a predefined sequence of service functions, realizing arbitrary network services through dynamic and efficient communication paths. Inspired by this concept, we propose an SFC-based architecture for Multi-hop Split Inference (MSI), where split sub-models are interpreted as service functions and their composition forms a service chain representing the global model. By leveraging SFC, the proposed architectur…

A note on the recovery sequence in the double gradient model for phase transitions
Jakob Deutsch
https://arxiv.org/abs/2510.01893 https://arxiv.org/pdf/251…

A note on the recovery sequence in the double gradient model for phase transitions
We investigate the $\limsup$ inequality in the double gradient model for phase transitions governed by a Modica--Mortola functional with a double-well potential in two dimensions. Specifically, we consider energy functionals of the form \[ E_\varepsilon(u, Ω) = \int_Ω\left( \frac{1}{\varepsilon} W(\nabla u) + \varepsilon |\nabla^2 u|^2 \right) dx \] for maps $ u \in H^2(Ω; \mathbb{R}^2) $, where $ W $ vanishes only at two wells. Assuming a bound on the optimal profile constant -- n…

Voter Model stability with respect to conservative noises
Gideon Amir, Omer Angel, Rangel Baldasso, Daniel de la Riva
https://arxiv.org/abs/2509.02717 https://

Voter Model stability with respect to conservative noises
The notions of noise sensitivity and stability were recently extended for the voter model. In this model, the vertices of a graph have opinions that are updated by uniformly selecting edges. We further extend stability results to different classes of perturbations. We consider two different types of noise: in the first one, an exclusion process is performed on the edge selections, while in the second, independent Brownian motions are applied to such a sequence. In both cases, we prove stability…

Pendulum Model of Spiking Neurons
Joy Bose
https://arxiv.org/abs/2507.22146 https://arxiv.org/pdf/2507.22146

Pendulum Model of Spiking Neurons
We propose a biologically inspired model of spiking neurons based on the dynamics of a damped, driven pendulum. Unlike traditional models such as the Leaky Integrate-and-Fire (LIF) neurons, the pendulum neuron incorporates second-order, nonlinear dynamics that naturally give rise to oscillatory behavior and phase-based spike encoding. This model captures richer temporal features and supports timing-sensitive computations critical for sequence processing and symbolic learning. We present an anal…

HITSZ's End-To-End Speech Translation Systems Combining Sequence-to-Sequence Auto Speech Recognition Model and Indic Large Language Model for IWSLT 2025 in Indic Track
Xuchen Wei, Yangxin Wu, Yaoyin Zhang, Henglyu Liu, Kehai Chen, Xuefeng Bai, Min Zhang
https://arxiv.org/abs/2507.19616

HITSZ's End-To-End Speech Translation Systems Combining Sequence-to-Sequence Auto Speech Recognition Model and Indic Large Language Model for IWSLT 2025 in Indic Track
This paper presents HITSZ's submission for the IWSLT 2025 Indic track, focusing on speech-to-text translation (ST) for English-to-Indic and Indic-to-English language pairs. To enhance translation quality in this low-resource scenario, we propose an end-to-end system integrating the pre-trained Whisper automated speech recognition (ASR) model with Krutrim, an Indic-specialized large language model (LLM). Experimental results demonstrate that our end-to-end system achieved average BLEU scores of …

Learning What's Missing: Attention Dispersion and EMA Stabilization in Length Generalization
P\'al Zs\'amboki, Benjamin Levi, David Ansel Josef Smith, Mitansh Kagalwala, Arlington Kell, Samuel Liechty, Cong Wang
https://arxiv.org/abs/2510.08341

Learning What's Missing: Attention Dispersion and EMA Stabilization in Length Generalization
We study length generalization in transformers through the set complement task, where a model must predict a uniform distribution over tokens absent from an input sequence -- an ability central to board-game style reasoning. Our main theoretical result establishes two statements. First, we prove tight bounds on embedding and value dimensions for single-layer attention-only transformers. Second, we show that if such a model achieves balanced logit displacement at lengths 1 and 2, then it must ge…

Dynamical Masses for 23 Pre-Main Sequence Stars in Upper Scorpius: A Critical Test of Stellar Evolutionary Models
A. P. M. Towner, J. A. Eisner, P. D. Sheehan, L. A. Hillenbrand, Y. -L. Wu
https://arxiv.org/abs/2509.23001

Dynamical Masses for 23 Pre-Main Sequence Stars in Upper Scorpius: A Critical Test of Stellar Evolutionary Models
We present dynamical masses for 23 pre-main sequence K- and M-type stars in the Upper Scorpius star-forming region. These masses were derived from the Keplerian rotation of CO disk gas using the MCMC radiative-transfer package pdspy and a flared-disk model with 15 free parameters. We compare our dynamical masses to those derived from five pre-main sequence evolutionary models, and find that most models consistently underestimate stellar mass by $\gtrsim$25%. Models with updated treatment of ste…

Modelling birdsong transmission with methods from molecular sequence analysis
Anthony Kwong, Mark Muldoon
https://arxiv.org/abs/2508.20833 https://arxiv.or…

Modelling birdsong transmission with methods from molecular sequence analysis
In many species of songbirds, juvenile males learn their songs from adult male tutors. In this paper we formulate a novel Markov model for birdsong transmission developed by analogy with models used in biological sequence analysis. We fit the model using the recently developed Interacting Particle Langevin Algorithm (IPLA) of Akyildiz et al. (arXiv:2303.13429) and analyse a collection of songs from Java sparrows (Lonchura oryzivora) originally recorded and studied by Masayo Soma and her collabo…

Quantum-Enhanced Natural Language Generation: A Multi-Model Framework with Hybrid Quantum-Classical Architectures
Chi-Sheng Chen, En-Jui Kuo
https://arxiv.org/abs/2508.21332 htt…

Quantum-Enhanced Natural Language Generation: A Multi-Model Framework with Hybrid Quantum-Classical Architectures
This paper presents a comprehensive evaluation of quantum text generation models against traditional Transformer/MLP architectures, addressing the growing interest in quantum computing applications for natural language processing. We conduct systematic experiments comparing five distinct models: Transformer (baseline), Quantum Kernel Self-Attention Network (QKSAN), Quantum RWKV (QRWKV), and Quantum Attention Sequence Architecture (QASA) across five diverse datasets including simple sentences, s…

To Sink or Not to Sink: Visual Information Pathways in Large Vision-Language Models
Jiayun Luo, Wan-Cyuan Fan, Lyuyang Wang, Xiangteng He, Tanzila Rahman, Purang Abolmaesumi, Leonid Sigal
https://arxiv.org/abs/2510.08510

To Sink or Not to Sink: Visual Information Pathways in Large Vision-Language Models
Large Vision Language Models (LVLMs) have recently emerged as powerful architectures capable of understanding and reasoning over both visual and textual information. These models typically rely on two key components: a Vision Transformer (ViT) and a Large Language Model (LLM). ViT encodes visual content into a sequence of image tokens and serves as the perceptual front-end -- the eyes of the model. In contrast, the LLM interprets these tokens to perform high-level reasoning, generates responses…

Gaussian Sequence Model: Sample Complexities of Testing, Estimation and LFHT
Zeyu Jia, Yury Polyanskiy
https://arxiv.org/abs/2507.16734 https://

Gaussian Sequence Model: Sample Complexities of Testing, Estimation and LFHT
We study the Gaussian sequence model, i.e. $X \sim N(\mathbfθ, I_\infty)$, where $\mathbfθ \in Γ\subset \ell_2$ is assumed to be convex and compact. We show that goodness-of-fit testing sample complexity is lower bounded by the square-root of the estimation complexity, whenever $Γ$ is orthosymmetric. We show that the lower bound is tight when $Γ$ is also quadratically convex, thus significantly extending validity of the testing-estimation relationship from [GP24]. Using similar methods, we…

Some new results on $\Delta$-spaces
I. Juh\'asz, J. van Mill, L. Soukup, Z. Szentmikl\'ossy
https://arxiv.org/abs/2510.04242 https://arxiv.org/pdf/…

Some new results on $Δ$-spaces
A topological space $X$ is a $Δ$-space (or $X \in Δ$) if for any decreasing sequence $\{A_n : n < ω\}$ of subsets of $X$ with empty intersection there is a (decreasing) sequence $\{U_n : n < ω\}$ of open sets with empty intersection such that $A_n \subset U_n$ for all $n < ω$. In this note we prove the following results concerning $Δ$-spaces. 1) Every $T_3$ countably compact $Δ$-space is compact. 2) If there is a $T_1$ crowded Baire $Δ$-space then there is an inner model with a meas…

Zonotope-Based Elastic Tube Model Predictive Control
Sabin Diaconescu, Florin Stoican, Bogdan D. Ciubotaru, Sorin Olaru
https://arxiv.org/abs/2509.19824 https://

Zonotope-Based Elastic Tube Model Predictive Control
Tube-based Model Predictive Control (MPC) is a widely adopted robust control framework for constrained linear systems under additive disturbance. The paper is focused on reducing the numerical complexity associated with the tube parameterization, described as a sequence of elastically-scaled zonotopic sets. A new class of scaled-zonotope inclusion conditions is proposed, alleviating the need for a priori specification of certain set-containment constraints and achieving significant reductions i…

ALMA Reveals an Eccentricity Gradient in the Fomalhaut Debris Disk
Joshua B. Lovell, Elliot M. Lynch, Jay Chittidi, Antranik A. Sefilian, Sean M. Andrews, Grant M. Kennedy, Meredith MacGregor, David J. Wilner, Mark C. Wyatt
https://arxiv.org/abs/2509.02884

ALMA Reveals an Eccentricity Gradient in the Fomalhaut Debris Disk
We present evidence of a negative eccentricity gradient in the debris disk of the nearby A-type main sequence star, Fomalhaut. Fitting to the high-resolution, archival ALMA 1.32 mm continuum data for Fomalhaut (with a synthesised angular resolution of $0.76{\times}0.55''$; 4-6\,au), we present a model that describes the bulk properties of the disk (semi-major axis, width, and geometry) and its asymmetric morphology. The best-fit model incorporates a forced eccentricity gradient that varies with…

Evolutionary Links: From Gaia Neutron Star Binaries to Pulsar White Dwarf Endpoints
Debatri Chattopadhyay, Kyle A. Rocha, Seth Gossage, Vicky Kalogera
https://arxiv.org/abs/2510.11828

Evolutionary Links: From Gaia Neutron Star Binaries to Pulsar White Dwarf Endpoints
The discovery of wide, eccentric Gaia neutron stars (NSs) in binaries with still evolving (likely main sequence) companions offers a new probe of mass transfer and pulsar recycling beyond the compact-binary regime. We model their origins and fates using population synthesis with POSYDON and detailed binary evolution with MESA, contrasting two limiting prescriptions at Roche-lobe overflow (RLOF): enforced circularization versus explicitly eccentric mass transfer. Our MESA setups include updated …

Decision Procedure for A Theory of String Sequences
Denghang Hu, Taolue Chen, Philipp R\"ummer, Fu Song, Zhilin Wu
https://arxiv.org/abs/2509.00948 https://

Decision Procedure for A Theory of String Sequences
The theory of sequences, supported by many SMT solvers, can model program data types including bounded arrays and lists. Sequences are parameterized by the element data type and provide operations such as accessing elements, concatenation, forming sub-sequences and updating elements. Strings and sequences are intimately related; many operations, e.g., matching a string according to a regular expression, splitting strings, or joining strings in a sequence, are frequently used in string-manipulat…

Fairness in Repeated Matching: A Maximin Perspective
Eugene Lim, Tzeh Yuan Neoh, Nicholas Teh
https://arxiv.org/abs/2510.04624 https://arxiv.org/pdf/2510.0…

Fairness in Repeated Matching: A Maximin Perspective
We study a sequential decision-making model where a set of items is repeatedly matched to the same set of agents over multiple rounds. The objective is to determine a sequence of matchings that either maximizes the utility of the least advantaged agent at the end of all rounds (optimal) or at the end of every individual round (anytime optimal). We investigate the computational challenges associated with finding (anytime) optimal outcomes and demonstrate that these problems are generally computa…

LSMTCR: A Scalable Multi-Architecture Model for Epitope-Specific T Cell Receptor de novo Design
Ruihao Zhang, Xiao Liu
https://arxiv.org/abs/2509.07627 https://

LSMTCR: A Scalable Multi-Architecture Model for Epitope-Specific T Cell Receptor de novo Design
Designing full-length, epitope-specific TCR α\b{eta} remains challenging due to vast sequence space, data biases and incomplete modeling of immunogenetic constraints. We present LSMTCR, a scalable multi-architecture framework that separates specificity from constraint learning to enable de novo, epitope-conditioned generation of paired, full-length TCRs. A diffusion-enhanced BERT encoder learns time-conditioned epitope representations; conditional GPT decoders, pretrained on CDR3\b{eta} and tr…

Latent-X: An Atom-level Frontier Model for De Novo Protein Binder Design
Latent Labs Team, Alex Bridgland, Jonathan Crabb\'e, Henry Kenlay, Daniella Pretorius, Sebastian M. Schmon, Agrin Hilmkil, Rebecca Bartke-Croughan, Robin Rombach, Michael Flashman, Tomas Matteson, Simon Mathis, Alexander W. R. Nelson, David Yuan, Annette Obika, Simon A. A. Kohl
https:/…

Latent-X: An Atom-level Frontier Model for De Novo Protein Binder Design
Traditional drug discovery relies on rounds of screening millions of candidate molecules with low success rates, making drug discovery time and resource intensive. To overcome this screening bottleneck, we introduce Latent-X, an all-atom protein design model that enables a new paradigm of precision AI design. Given a target protein epitope, Latent-X jointly generates the all atom structure and sequence of the protein binder and target, directly modelling the non-covalent interactions essential …

DNABERT-2: Fine-Tuning a Genomic Language Model for Colorectal Gene Enhancer Classification
Darren King, Yaser Atlasi, Gholamreza Rafiee
https://arxiv.org/abs/2509.25274 https:/…

DNABERT-2: Fine-Tuning a Genomic Language Model for Colorectal Gene Enhancer Classification
Gene enhancers control when and where genes switch on, yet their sequence diversity and tissue specificity make them hard to pinpoint in colorectal cancer. We take a sequence-only route and fine-tune DNABERT-2, a transformer genomic language model that uses byte-pair encoding to learn variable-length tokens from DNA. Using assays curated via the Johnston Cancer Research Centre at Queen's University Belfast, we assembled a balanced corpus of 2.34 million 1 kb enhancer sequences, applied summit-c…

Thinkquel: A Model Dedicated to Text-to-dbt Using Synthetic Data and a Span-Aware Objective
Anni Li, Aria Attar, Paul Dong
https://arxiv.org/abs/2510.00186 https://

Thinkquel: A Model Dedicated to Text-to-dbt Using Synthetic Data and a Span-Aware Objective
Transforming natural-language requests into reliable, production-ready data transformations remains challenging: correctness depends on precise schema linking and warehouse-specific SQL dialects, while the strongest supervision available during training--execution success and result matching--are provided only at the sequence level. At the same time, assembling large, execution-validated corpora is costly, and token-level objectives misalign with these global signals, yielding unstable optimiza…

DeRainMamba: A Frequency-Aware State Space Model with Detail Enhancement for Image Deraining
Zhiliang Zhu, Tao Zeng, Tao Yang, Guoliang Luo, Jiyong Zeng
https://arxiv.org/abs/2510.06746

DeRainMamba: A Frequency-Aware State Space Model with Detail Enhancement for Image Deraining
Image deraining is crucial for improving visual quality and supporting reliable downstream vision tasks. Although Mamba-based models provide efficient sequence modeling, their limited ability to capture fine-grained details and lack of frequency-domain awareness restrict further improvements. To address these issues, we propose DeRainMamba, which integrates a Frequency-Aware State-Space Module (FASSM) and Multi-Directional Perception Convolution (MDPConv). FASSM leverages Fourier transform to d…

PagedEviction: Structured Block-wise KV Cache Pruning for Efficient Large Language Model Inference
Krishna Teja Chitty-Venkata, Jie Ye, Xian-He Sun, Anthony Kougkas, Murali Emani, Venkatram Vishwanath, Bogdan Nicolae
https://arxiv.org/abs/2509.04377

PagedEviction: Structured Block-wise KV Cache Pruning for Efficient Large Language Model Inference
KV caching significantly improves the efficiency of Large Language Model (LLM) inference by storing attention states from previously processed tokens, enabling faster generation of subsequent tokens. However, as sequence length increases, the KV cache quickly becomes a major memory bottleneck. To address this, we propose PagedEviction, a novel fine-grained, structured KV cache pruning strategy that enhances the memory efficiency of vLLM's PagedAttention. Unlike existing approaches that rely on …

An Optimal Sorting Algorithm for Persistent Random Comparison Faults
Barbara Geissmann, Stefano Leucci, Chih-Hung Liu, Paolo Penna
https://arxiv.org/abs/2508.19785 https://

An Optimal Sorting Algorithm for Persistent Random Comparison Faults
We consider the problem of sorting $n$ elements subject to persistent random comparison errors. In this problem, each comparison between two elements can be wrong with some fixed (small) probability $p$, and comparing the same pair of elements multiple times always yields the same result. Sorting perfectly in this model is impossible, and the objective is to minimize the dislocation of each element in the output sequence, i.e., the difference between its position in the sequence and its true ra…

SpecMamba: Accelerating Mamba Inference on FPGA with Speculative Decoding
Linfeng Zhong, Songqiang Xu, Huifeng Wen, Tong Xie, Qingyu Guo, Yuan Wang, Meng Li
https://arxiv.org/abs/2509.19873

SpecMamba: Accelerating Mamba Inference on FPGA with Speculative Decoding
The growing demand for efficient long-sequence modeling on edge devices has propelled widespread adoption of State Space Models (SSMs) like Mamba, due to their superior computational efficiency and scalability. As its autoregressive generation process remains memory-bound, speculative decoding has been proposed that incorporates draft model generation and target model verification. However, directly applying speculative decoding to SSMs faces three key challenges: (1) hidden state backtracking …

INSIGHT: INference-time Sequence Introspection for Generating Help Triggers in Vision-Language-Action Models
Ulas Berk Karli, Ziyao Shangguan, Tesca FItzgerald
https://arxiv.org/abs/2510.01389

INSIGHT: INference-time Sequence Introspection for Generating Help Triggers in Vision-Language-Action Models
Recent Vision-Language-Action (VLA) models show strong generalization capabilities, yet they lack introspective mechanisms for anticipating failures and requesting help from a human supervisor. We present \textbf{INSIGHT}, a learning framework for leveraging token-level uncertainty signals to predict when a VLA should request help. Using $π_0$-FAST as the underlying model, we extract per-token \emph{entropy}, \emph{log-probability}, and Dirichlet-based estimates of \emph{aleatoric and epistemi…

On a Neutral Host-Virus Model with Recombination
Raphael Eichhorn, Cornelia Pokalyuk
https://arxiv.org/abs/2509.19253 https://arxiv.org/pdf/2509.19253

The Computational Complexity of Satisfiability in State Space Models
Eric Alsmann, Martin Lange
https://arxiv.org/abs/2508.18162 https://arxiv.org/pdf/2508…

The Computational Complexity of Satisfiability in State Space Models
We analyse the complexity of the satisfiability problem ssmSAT for State Space Models (SSM), which asks whether an input sequence can lead the model to an accepting configuration. We find that ssmSAT is undecidable in general, reflecting the computational power of SSM. Motivated by practical settings, we identify two natural restrictions under which ssmSAT becomes decidable and establish corresponding complexity bounds. First, for SSM with bounded context length, ssmSAT is NP-complete when the …

BurstEngine: an Efficient Distributed Framework for Training Transformers on Extremely Long Sequences of over 1M Tokens
Ao Sun, Weilin Zhao, Xu Han, Cheng Yang, Zhiyuan Liu, Chuan Shi, Maosong sun
https://arxiv.org/abs/2509.19836

BurstEngine: an Efficient Distributed Framework for Training Transformers on Extremely Long Sequences of over 1M Tokens
Existing methods for training LLMs on long-sequence data, such as Tensor Parallelism and Context Parallelism, exhibit low Model FLOPs Utilization as sequence lengths and number of GPUs increase, especially when sequence lengths exceed 1M tokens. To address these challenges, we propose BurstEngine, an efficient framework designed to train LLMs on long-sequence data. BurstEngine introduces BurstAttention, an optimized distributed attention with lower communication cost than RingAttention. BurstAt…

On the Theory of Continual Learning with Gradient Descent for Neural Networks
Hossein Taheri, Avishek Ghosh, Arya Mazumdar
https://arxiv.org/abs/2510.05573 https://

On the Theory of Continual Learning with Gradient Descent for Neural Networks
Continual learning, the ability of a model to adapt to an ongoing sequence of tasks without forgetting the earlier ones, is a central goal of artificial intelligence. To shed light on its underlying mechanisms, we analyze the limitations of continual learning in a tractable yet representative setting. In particular, we study one-hidden-layer quadratic neural networks trained by gradient descent on an XOR cluster dataset with Gaussian noise, where different tasks correspond to different clusters…

On Recurrence of the Infinite Server Queue
Sergey Foss, Peter Glynn
https://arxiv.org/abs/2509.06075 https://arxiv.org/pdf/2509.06075

On Recurrence of the Infinite Server Queue
This paper concerns the recurrence structure of the infinite server queue, as viewed through the prism of the maximum dater sequence, namely the time to drain the current work in the system as seen at arrival epochs. Despite the importance of this model in queueing theory, we are aware of no complete analysis of the stability behavior of this model, especially in settings in which either or both the inter arrival and service time distributions have infinite mean. In this paper, we fully develop…

Integrating LLM-Derived Multi-Semantic Intent into Graph Model for Session-based Recommendation
Shuo Zhang, Xiao Li, Jiayi Wu, Fan Yang, Xiang Li, Ming Gao
https://arxiv.org/abs/2507.20147

Integrating LLM-Derived Multi-Semantic Intent into Graph Model for Session-based Recommendation
Session-based recommendation (SBR) is mainly based on anonymous user interaction sequences to recommend the items that the next user is most likely to click. Currently, the most popular and high-performing SBR methods primarily leverage graph neural networks (GNNs), which model session sequences as graph-structured data to effectively capture user intent. However, most GNNs-based SBR methods primarily focus on modeling the ID sequence information of session sequences, while neglecting the rich …

Quasar Main Sequence unfolded by 2.5D FRADO (Natural expression of Eddington ratio, black hole mass, and inclination)
M. H. Naddaf, M. L. Mart\'inez-Aldama, P. Marziani, B. Czerny, D. Hutsem\'ekers
https://arxiv.org/abs/2510.01092

Quasar Main Sequence unfolded by 2.5D FRADO (Natural expression of Eddington ratio, black hole mass, and inclination)
The quasar main sequence (QMS), characterized by the Eigenvector 1 (EV1), serves as a unifying framework for classifying type-1 active galactic nuclei (AGNs) based on their diverse spectral properties. Although a fully self-consistent physical interpretation has long been lacking, our physically motivated 2.5D FRADO (Failed Radiatively Accelerated Dusty Outflow) model naturally predicts that the Eddington ratio ($\dot{m}$) is the primary physical driver of the QMS, with black hole mass ($M_{\rm…

Transforming Questions and Documents for Semantically Aligned Retrieval-Augmented Generation
Seokgi Lee
https://arxiv.org/abs/2508.09755 https://arxiv.org/…

Transforming Questions and Documents for Semantically Aligned Retrieval-Augmented Generation
We introduce a novel retrieval-augmented generation (RAG) framework tailored for multihop question answering. First, our system uses large language model (LLM) to decompose complex multihop questions into a sequence of single-hop subquestions that guide document retrieval. This decomposition mitigates the ambiguity inherent in multi-hop queries by clearly targeting distinct knowledge facets. Second, instead of embedding raw or chunked documents directly, we generate answerable questions from ea…

Attention as an Adaptive Filter
Peter Racioppo
https://arxiv.org/abs/2509.04154 https://arxiv.org/pdf/2509.04154 …

Attention as an Adaptive Filter
We introduce Adaptive Filter Attention (AFA), a novel attention mechanism that incorporates a learnable dynamics model directly into the computation of attention weights. Rather than comparing queries and keys directly, we model the input sequence as discrete observations of a linear stochastic differential equation (SDE). By imposing a linear dynamics model with simultaneously diagonalizable state matrices and noise covariances, we can make use of a closed-form solution to the differential Lya…

Controllable Singing Voice Synthesis using Phoneme-Level Energy Sequence
Yerin Ryu, Inseop Shin, Chanwoo Kim
https://arxiv.org/abs/2509.07038 https://arxiv…

Controllable Singing Voice Synthesis using Phoneme-Level Energy Sequence
Controllable Singing Voice Synthesis (SVS) aims to generate expressive singing voices reflecting user intent. While recent SVS systems achieve high audio quality, most rely on probabilistic modeling, limiting precise control over attributes such as dynamics. We address this by focusing on dynamic control--temporal loudness variation essential for musical expressiveness--and explicitly condition the SVS model on energy sequences extracted from ground-truth spectrograms, reducing annotation costs…

Chain-of-Trigger: An Agentic Backdoor that Paradoxically Enhances Agentic Robustness
Jiyang Qiu, Xinbei Ma, Yunqing Xu, Zhuosheng Zhang, Hai Zhao
https://arxiv.org/abs/2510.08238

Chain-of-Trigger: An Agentic Backdoor that Paradoxically Enhances Agentic Robustness
The rapid deployment of large language model (LLM)-based agents in real-world applications has raised serious concerns about their trustworthiness. In this work, we reveal the security and robustness vulnerabilities of these agents through backdoor attacks. Distinct from traditional backdoors limited to single-step control, we propose the Chain-of-Trigger Backdoor (CoTri), a multi-step backdoor attack designed for long-horizon agentic control. CoTri relies on an ordered sequence. It starts with…

MESA Isochrones and Stellar Tracks (MIST) III. The White Dwarf Cooling Sequence
Evan B. Bauer, Aaron Dotter, Charlie Conroy, Tim Cunningham, Minjung Park, Pier-Emmanuel Tremblay
https://arxiv.org/abs/2509.21717

MESA Isochrones and Stellar Tracks (MIST) III. The White Dwarf Cooling Sequence
We present a substantial update to the MESA Isochrones and Stellar Tracks (MIST) library, extending the MIST model grids and isochrones down the white dwarf (WD) cooling sequence with realistic physics for WD cooling timescales. This work provides a large grid of MESA models for carbon-oxygen core WDs with hydrogen atmospheres (spectral type DA/DC), descended from full prior stellar evolution calculations. The model tracks, isochrones, and WD cooling timescale contours are available on the MIST…

Physicochemically Informed Dual-Conditioned Generative Model of T-Cell Receptor Variable Regions for Cellular Therapy
Jiahao Ma, Hongzong Li, Ye-Fan Hu, Jian-Dong Huang
https://arxiv.org/abs/2510.05747

Physicochemically Informed Dual-Conditioned Generative Model of T-Cell Receptor Variable Regions for Cellular Therapy
Physicochemically informed biological sequence generation has the potential to accelerate computer-aided cellular therapy, yet current models fail to \emph{jointly} ensure novelty, diversity, and biophysical plausibility when designing variable regions of T-cell receptors (TCRs). We present \textbf{PhysicoGPTCR}, a large generative protein Transformer that is \emph{dual-conditioned} on peptide and HLA context and trained to autoregressively synthesise TCR sequences while embedding residue-level…

Reducing Quantum Circuit Synthesis to #SAT
Dekel Zak, Jingyi Mei, Jean-Marie Lagniez, Alfons Laarman
https://arxiv.org/abs/2508.00416

Reducing Quantum Circuit Synthesis to #SAT
Quantum circuit synthesis is the task of decomposing a given quantum operator into a sequence of elementary quantum gates. Since the finite target gate set cannot exactly implement any given operator, approximation is often necessary. Model counting, or #SAT, has recently been demonstrated as a promising new approach for tackling core problems in quantum circuit analysis. In this work, we show for the first time that the universal quantum circuit synthesis problem can be reduced to maximum mode…

Zeppelin: Balancing Variable-length Workloads in Data Parallel Large Model Training
Chang Chen, Tiancheng Chen, Jiangfei Duan, Qianchao Zhu, Zerui Wang, Qinghao Hu, Peng Sun, Xiuhong Li, Chao Yang, Torsten Hoefler
https://arxiv.org/abs/2509.21841

Zeppelin: Balancing Variable-length Workloads in Data Parallel Large Model Training
Training large language models (LLMs) with increasingly long and varying sequence lengths introduces severe load imbalance challenges in large-scale data-parallel training. Recent frameworks attempt to mitigate these issues through data reorganization or hybrid parallel strategies. However, they often overlook how computational and communication costs scale with sequence length, resulting in suboptimal performance. We identify three critical challenges: (1) varying computation-to-communication …

Why Stop at Words? Unveiling the Bigger Picture through Line-Level OCR
Shashank Vempati, Nishit Anand, Gaurav Talebailkar, Arpan Garai, Chetan Arora
https://arxiv.org/abs/2508.21693

Why Stop at Words? Unveiling the Bigger Picture through Line-Level OCR
Conventional optical character recognition (OCR) techniques segmented each character and then recognized. This made them prone to error in character segmentation, and devoid of context to exploit language models. Advances in sequence to sequence translation in last decade led to modern techniques first detecting words and then inputting one word at a time to a model to directly output full words as sequence of characters. This allowed better utilization of language models and bypass error-prone…

Guiding an Automatic Speech Recognition Decoder Using Large Language Models
Eyal Cohen (Technion - Israel Institute of Technology), Bhiksha Raj (Carnegie Mellon University), Joseph Keshet (Technion - Israel Institute of Technology)
https://arxiv.org/abs/2508.02228

Guiding an Automatic Speech Recognition Decoder Using Large Language Models
Automatic Speech Recognition (ASR) consists of an acoustic model (AM) and a language model (LM). The AM estimates the probability of an acoustic signal based on a sequence of linguistic units, typically phones, characters, or tokens, while the LM assesses the likelihood of a specific sequence of words or tokens. Although Large Language Models (LLMs) have demonstrated significant potential across various tasks, integrating them into ASR remains an open challenge. By decomposing the maximum a pos…

InstructPLM-mu: 1-Hour Fine-Tuning of ESM2 Beats ESM3 in Protein Mutation Predictions
Junde Xu, Yapin Shi, Lijun Lang, Taoyong Cui, Zhiming Zhang, Guangyong Chen, Jiezhong Qiu, Pheng-Ann Heng
https://arxiv.org/abs/2510.03370

InstructPLM-mu: 1-Hour Fine-Tuning of ESM2 Beats ESM3 in Protein Mutation Predictions
Multimodal protein language models deliver strong performance on mutation-effect prediction, but training such models from scratch demands substantial computational resources. In this paper, we propose a fine-tuning framework called InstructPLM-mu and try to answer a question: \textit{Can multimodal fine-tuning of a pretrained, sequence-only protein language model match the performance of models trained end-to-end? } Surprisingly, our experiments show that fine-tuning ESM2 with structural input…

General Proximal Quasi-Newton Methods based on model functions for nonsmooth nonconvex problems
Xiaoxi Jia, Peter Ochs
https://arxiv.org/abs/2507.18363 https://

General Proximal Quasi-Newton Methods based on model functions for nonsmooth nonconvex problems
In this manuscript, we propose a general proximal quasi-Newton method tailored for nonconvex and nonsmooth optimization problems, where we do not require the sequence of the variable metric (or Hessian approximation) to be uniformly bounded as a prerequisite, instead, the variable metric is updated by a continuous matrix generator. From the respective of the algorithm, the objective function is approximated by the so-called local model function and subproblems aim to exploit the proximal point(…

Next Point-of-interest (POI) Recommendation Model Based on Multi-modal Spatio-temporal Context Feature Embedding
Lingyu Zhang, Guobin Wu, Yan Wang, Pengfei Xu, Jian Liang, Xuan Song, Yunhai Wang
https://arxiv.org/abs/2509.22661

Next Point-of-interest (POI) Recommendation Model Based on Multi-modal Spatio-temporal Context Feature Embedding
The next Point-of-interest (POI) recommendation is mainly based on sequential traffic information to predict the user's next boarding point location. This is a highly regarded and widely applied research task in the field of intelligent transportation, and there have been many research results to date. Traditional POI prediction models primarily rely on short-term traffic sequence information, often neglecting both long-term and short-term preference data, as well as crucial spatiotemporal cont…

Lossless Vocabulary Reduction for Auto-Regressive Language Models
Daiki Chijiwa, Taku Hasegawa, Kyosuke Nishida, Shin'ya Yamaguchi, Tomoya Ohba, Tamao Sakao, Susumu Takeuchi
https://arxiv.org/abs/2510.08102

Lossless Vocabulary Reduction for Auto-Regressive Language Models
Tokenization -- the process of decomposing a given text into a sequence of subwords called tokens -- is one of the key components in the development of language models. Particularly, auto-regressive language models generate texts token by token, i.e., by predicting the next-token distribution given the previous ones, and thus tokenization directly affects their efficiency in text generation. Since each language model has their own vocabulary as a set of possible tokens, they struggle to coopera…

Uncertainty Under the Curve: A Sequence-Level Entropy Area Metric for Reasoning LLM
Yongfu Zhu, Lin Sun, Guangxiang Zhao, Weihong Lin, Xiangzheng Zhang
https://arxiv.org/abs/2508.20384

Uncertainty Under the Curve: A Sequence-Level Entropy Area Metric for Reasoning LLM
In this work, we introduce Entropy Area Score (EAS), a simple yet effective metric to quantify uncertainty in the answer generation process of reasoning large language models (LLMs). EAS requires neither external models nor repeated sampling, it integrates token-level predictive entropy from the model itself to capture the evolution of uncertainty during generation. Empirical results show that EAS is strongly correlated with answer entropy across models and datasets. In training data selection,…

Is Sequence Information All You Need for Bayesian Optimization of Antibodies?
Sebastian W. Ober, Calvin McCarter, Aniruddh Raghu, Yucen Lily Li, Alan N. Amin, Andrew Gordon Wilson, Hunter Elliott
https://arxiv.org/abs/2509.24933

Is Sequence Information All You Need for Bayesian Optimization of Antibodies?
Bayesian optimization is a natural candidate for the engineering of antibody therapeutic properties, which is often iterative and expensive. However, finding the optimal choice of surrogate model for optimization over the highly structured antibody space is difficult, and may differ depending on the property being optimized. Moreover, to the best of our knowledge, no prior works have attempted to incorporate structural information into antibody Bayesian optimization. In this work, we explore di…

Training chord recognition models on artificially generated audio
Martyna Majchrzak, Jacek Ma\'ndziuk
https://arxiv.org/abs/2508.05878 https://arxiv.or…

Training chord recognition models on artificially generated audio
One of the challenging problems in Music Information Retrieval is the acquisition of enough non-copyrighted audio recordings for model training and evaluation. This study compares two Transformer-based neural network models for chord sequence recognition in audio recordings and examines the effectiveness of using an artificially generated dataset for this purpose. The models are trained on various combinations of Artificial Audio Multitracks (AAM), Schubert's Winterreise Dataset, and the McGill…

Inductive Bias and Spectral Properties of Single-Head Attention in High Dimensions
Fabrizio Boncoraglio, Vittorio Erba, Emanuele Troiani, Florent Krzakala, Lenka Zdeborov\'a
https://arxiv.org/abs/2509.24914

Inductive Bias and Spectral Properties of Single-Head Attention in High Dimensions
We study empirical risk minimization in a single-head tied-attention layer trained on synthetic high-dimensional sequence tasks, given by the recently introduced attention-indexed model. Using tools from random matrix theory, spin-glass physics, and approximate message passing, we derive sharp asymptotics for training and test errors, locate interpolation and recovery thresholds, and characterize the limiting spectral distribution of the learned weights. Weight decay induces an implicit nuclear…

A HyperGraphMamba-Based Multichannel Adaptive Model for ncRNA Classification
Xin An, Ruijie Li, Qiao Ning, Hui Li, Qian Ma, Shikai Guo
https://arxiv.org/abs/2509.20240 https://

A HyperGraphMamba-Based Multichannel Adaptive Model for ncRNA Classification
Non-coding RNAs (ncRNAs) play pivotal roles in gene expression regulation and the pathogenesis of various diseases. Accurate classification of ncRNAs is essential for functional annotation and disease diagnosis. To address existing limitations in feature extraction depth and multimodal fusion, we propose HGMamba-ncRNA, a HyperGraphMamba-based multichannel adaptive model, which integrates sequence, secondary structure, and optionally available expression features of ncRNAs to enhance classificat…

Latent Thinking Optimization: Your Latent Reasoning Language Model Secretly Encodes Reward Signals in its Latent Thoughts
Hanwen Du, Yuxin Dong, Xia Ning
https://arxiv.org/abs/2509.26314

Latent Thinking Optimization: Your Latent Reasoning Language Model Secretly Encodes Reward Signals in its Latent Thoughts
Large Language Models (LLMs) excel at problem solving by generating chain of thoughts in natural language, but such verbal thinking is computationally costly and prone to overthinking. Recent work instead proposes a latent thinking architecture Huggin-3.5B, which represents intermediate reasoning steps as sequence of latent representations. However, latent thoughts lack interpretability and are difficult to supervise, raising concerns about the correctness and reliability of its latent thinking…

Enhancing compact convolutional transformers with super attention
Simpenzwe Honore Leandre, Natenaile Asmamaw Shiferaw, Dillip Rout
https://arxiv.org/abs/2508.18960 https://

Enhancing compact convolutional transformers with super attention
In this paper, we propose a vision model that adopts token mixing, sequence-pooling, and convolutional tokenizers to achieve state-of-the-art performance and efficient inference in fixed context-length tasks. In the CIFAR100 benchmark, our model significantly improves the baseline of the top 1% and top 5% validation accuracy from 36.50% to 46.29% and 66.33% to 76.31%, while being more efficient than the Scaled Dot Product Attention (SDPA) transformers when the context length is less than the em…

DMGIN: How Multimodal LLMs Enhance Large Recommendation Models for Lifelong User Post-click Behaviors
Zhuoxing Wei, Qingchen Xie, Qi Liu
https://arxiv.org/abs/2508.21801 https:/…

DMGIN: How Multimodal LLMs Enhance Large Recommendation Models for Lifelong User Post-click Behaviors
Modeling user interest based on lifelong user behavior sequences is crucial for enhancing Click-Through Rate (CTR) prediction. However, long post-click behavior sequences themselves pose severe performance issues: the sheer volume of data leads to high computational costs and inefficiencies in model training and inference. Traditional methods address this by introducing two-stage approaches, but this compromises model effectiveness due to incomplete utilization of the full sequence context. Mor…

M3DIS - A grid of 3D radiation-hydrodynamics stellar atmosphere models for stellar surveys. II. Carbon-enhanced metal-poor stars
Philipp Eitner, Maria Bergemann, Richard Hoppe, Nicholas Storm, Veronika Lipatova, Simon C. O. Glover, Ralf S. Klessen, {\AA}ke Nordlund, Andrius Popovas
https://arxiv.org/abs/2509.24555

M3DIS - A grid of 3D radiation-hydrodynamics stellar atmosphere models for stellar surveys. II. Carbon-enhanced metal-poor stars
Understanding the origin and evolution of carbon-enhanced metal-poor (CEMP) stars is key to tracing the Galaxy's early chemical enrichment. We investigate how realistic 3D radiation-hydrodynamics (RHD) model atmospheres affect carbon abundances in CEMP stars and implications for their classification and Galactic chemical evolution (GCE). We focus on biases from traditional 1D hydrostatic models. We use the M3DIS code to compute 3D RHD model atmospheres for main-sequence and sub-giant stars over…

PRISM-Consult: A Panel-of-Experts Architecture for Clinician-Aligned Diagnosis
Lionel Levine, John Santerre, Alexander S. Young, T. Barry Levine, Francis Campion, Majid Sarrafzadeh
https://arxiv.org/abs/2510.01114

PRISM-Consult: A Panel-of-Experts Architecture for Clinician-Aligned Diagnosis
We present PRISM-Consult, a clinician-aligned panel-of-experts architecture that extends the compact PRISM sequence model into a routed family of domain specialists. Episodes are tokenized as structured clinical events; a light-weight router reads the first few tokens and dispatches to specialist models (Cardiac-Vascular, Pulmonary, Gastro-Oesophageal, Musculoskeletal, Psychogenic). Each specialist inherits PRISM's small transformer backbone and token template, enabling parameter efficiency and…

Refine Drugs, Don't Complete Them: Uniform-Source Discrete Flows for Fragment-Based Drug Discovery
Benno Kaech, Luis Wyss, Karsten Borgwardt, Gianvito Grasso
https://arxiv.org/abs/2509.26405

Refine Drugs, Don't Complete Them: Uniform-Source Discrete Flows for Fragment-Based Drug Discovery
We introduce InVirtuoGen, a discrete flow generative model for fragmented SMILES for de novo and fragment-constrained generation, and target-property/lead optimization of small molecules. The model learns to transform a uniform source over all possible tokens into the data distribution. Unlike masked models, its training loss accounts for predictions on all sequence positions at every denoising step, shifting the generation paradigm from completion to refinement, and decoupling the number of sa…

Hybrid Tokenization Strategy for DNA Language Model using Byte Pair Encoding and K-MER Methods
Ganesh Sapkota, Md Hasibur Rahman
https://arxiv.org/abs/2507.18570 https://…

Hybrid Tokenization Strategy for DNA Language Model using Byte Pair Encoding and K-MER Methods
This paper presents a novel hybrid tokenization strategy that enhances the performance of DNA Language Models (DLMs) by combining 6-mer tokenization with Byte Pair Encoding (BPE-600). Traditional k-mer tokenization is effective at capturing local DNA sequence structures but often faces challenges, including uneven token distribution and a limited understanding of global sequence context. To address these limitations, we propose merging unique 6mer tokens with optimally selected BPE tokens gener…

Amadeus: Autoregressive Model with Bidirectional Attribute Modelling for Symbolic Music
Hongju Su, Ke Li, Lan Yang, Honggang Zhang, Yi-Zhe Song
https://arxiv.org/abs/2508.20665 …

Amadeus: Autoregressive Model with Bidirectional Attribute Modelling for Symbolic Music
Existing state-of-the-art symbolic music generation models predominantly adopt autoregressive or hierarchical autoregressive architectures, modelling symbolic music as a sequence of attribute tokens with unidirectional temporal dependencies, under the assumption of a fixed, strict dependency structure among these attributes. However, we observe that using different attributes as the initial token in these models leads to comparable performance. This suggests that the attributes of a musical not…

FastVGGT: Training-Free Acceleration of Visual Geometry Transformer
You Shen, Zhipeng Zhang, Yansong Qu, Liujuan Cao
https://arxiv.org/abs/2509.02560 https://

FastVGGT: Training-Free Acceleration of Visual Geometry Transformer
Foundation models for 3D vision have recently demonstrated remarkable capabilities in 3D perception. However, scaling these models to long-sequence image inputs remains a significant challenge due to inference-time inefficiency. In this work, we present a detailed analysis of VGGT, a state-of-the-art feed-forward visual geometry model and identify its primary bottleneck. Visualization further reveals a token collapse phenomenon in the attention maps. Motivated by these findings, we explore the …

PRISM-Consult: A Panel-of-Experts Architecture for Clinician-Aligned Diagnosis
Lionel Levine, John Santerre, Alexander S. Young, T. Barry Levine, Francis Campion, Majid Sarrafzadeh
https://arxiv.org/abs/2510.01114

PRISM-Consult: A Panel-of-Experts Architecture for Clinician-Aligned Diagnosis
We present PRISM-Consult, a clinician-aligned panel-of-experts architecture that extends the compact PRISM sequence model into a routed family of domain specialists. Episodes are tokenized as structured clinical events; a light-weight router reads the first few tokens and dispatches to specialist models (Cardiac-Vascular, Pulmonary, Gastro-Oesophageal, Musculoskeletal, Psychogenic). Each specialist inherits PRISM's small transformer backbone and token template, enabling parameter efficiency and…

Spacetime-GR: A Spacetime-Aware Generative Model for Large Scale Online POI Recommendation
Haitao Lin, Zhen Yang, Jiawei Xue, Ziji Zhang, Luzhu Wang, Yikun Gu, Yao Xu, Xin Li
https://arxiv.org/abs/2508.16126

Spacetime-GR: A Spacetime-Aware Generative Model for Large Scale Online POI Recommendation
Building upon the strong sequence modeling capability, Generative Recommendation (GR) has gradually assumed a dominant position in the application of recommendation tasks (e.g., video and product recommendation). However, the application of Generative Recommendation in Point-of-Interest (POI) recommendation, where user preferences are significantly affected by spatiotemporal variations, remains a challenging open problem. In this paper, we propose Spacetime-GR, the first spacetime-aware generat…

Rethinking the long-range dependency in Mamba/SSM and transformer models
Cong Ma, Kayvan Najarian
https://arxiv.org/abs/2509.04226 https://arxiv.org/pdf/25…

Rethinking the long-range dependency in Mamba/SSM and transformer models
Long-range dependency is one of the most desired properties of recent sequence models such as state-space models (particularly Mamba) and transformer models. New model architectures are being actively developed and benchmarked for prediction tasks requiring long-range dependency. However, the capability of modeling long-range dependencies of these models has not been investigated from a theoretical perspective, which hinders a systematic improvement on this aspect. In this work, we mathematical…

SciReasoner: Laying the Scientific Reasoning Ground Across Disciplines
Yizhou Wang, Chen Tang, Han Deng, Jiabei Xiao, Jiaqi Liu, Jianyu Wu, Jun Yao, Pengze Li, Encheng Su, Lintao Wang, Guohang Zhuang, Yuchen Ren, Ben Fei, Ming Hu, Xin Chen, Dongzhan Zhou, Junjun He, Xiangyu Yue, Zhenfei Yin, Jiamin Wu, Qihao Zheng, Yuhao Zhou, Huihui Xu, Chenglong Ma, Yan Lu, Wenlong Zhang, Chunfeng Song, Philip Torr, Shixiang Tang, Xinzhu Ma, Wanli Ouyang, Lei Bai

SciReasoner: Laying the Scientific Reasoning Ground Across Disciplines
We present a scientific reasoning foundation model that aligns natural language with heterogeneous scientific representations. The model is pretrained on a 206B-token corpus spanning scientific text, pure sequences, and sequence-text pairs, then aligned via SFT on 40M instructions, annealed cold-start bootstrapping to elicit long-form chain-of-thought, and reinforcement learning with task-specific reward shaping, which instills deliberate scientific reasoning. It supports four capability famili…

Rethinking Transformer Connectivity: TLinFormer, A Path to Exact, Full Context-Aware Linear Attention
Zhongpan Tang
https://arxiv.org/abs/2508.20407 https://

Rethinking Transformer Connectivity: TLinFormer, A Path to Exact, Full Context-Aware Linear Attention
The Transformer architecture has become a cornerstone of modern artificial intelligence, but its core self-attention mechanism suffers from a complexity bottleneck that scales quadratically with sequence length, severely limiting its application in long-sequence tasks. To address this challenge, existing linear attention methods typically sacrifice model performance by relying on data-agnostic kernel approximations or restrictive context selection. This paper returns to the first principles of …

LINKER: Learning Interactions Between Functional Groups and Residues With Chemical Knowledge-Enhanced Reasoning and Explainability
Phuc Pham, Viet Thanh Duy Nguyen, Truong-Son Hy
https://arxiv.org/abs/2509.03425

LINKER: Learning Interactions Between Functional Groups and Residues With Chemical Knowledge-Enhanced Reasoning and Explainability
Accurate identification of interactions between protein residues and ligand functional groups is essential to understand molecular recognition and guide rational drug design. Existing deep learning approaches for protein-ligand interpretability often rely on 3D structural input or use distance-based contact labels, limiting both their applicability and biological relevance. We introduce LINKER, the first sequence-based model to predict residue-functional group interactions in terms of biologica…

High-Dimensional Analysis of Single-Layer Attention for Sparse-Token Classification
Nicholas Barnfield, Hugo Cui, Yue M. Lu
https://arxiv.org/abs/2509.25153 https://

High-Dimensional Analysis of Single-Layer Attention for Sparse-Token Classification
When and how can an attention mechanism learn to selectively attend to informative tokens, thereby enabling detection of weak, rare, and sparsely located features? We address these questions theoretically in a sparse-token classification model in which positive samples embed a weak signal vector in a randomly chosen subset of tokens, whereas negative samples are pure noise. In the long-sequence limit, we show that a simple single-layer attention classifier can in principle achieve vanishing tes…

DATS: Distance-Aware Temperature Scaling for Calibrated Class-Incremental Learning
Giuseppe Serra, Florian Buettner
https://arxiv.org/abs/2509.21161 https://

DATS: Distance-Aware Temperature Scaling for Calibrated Class-Incremental Learning
Continual Learning (CL) is recently gaining increasing attention for its ability to enable a single model to learn incrementally from a sequence of new classes. In this scenario, it is important to keep consistent predictive performance across all the classes and prevent the so-called Catastrophic Forgetting (CF). However, in safety-critical applications, predictive performance alone is insufficient. Predictive models should also be able to reliably communicate their uncertainty in a calibrated…

HAEPO: History-Aggregated Exploratory Policy Optimization
Gaurish Trivedi, Alakh Sharma, Kartikey Singh Bhandari, Dhruv Kumar, Pratik Narang, Jagat Sesh Challa
https://arxiv.org/abs/2508.18884

HAEPO: History-Aggregated Exploratory Policy Optimization
Exploration is essential in modern learning, from reinforcement learning environments with small neural policies to large language models (LLMs). Existing work, such as DPO, leverages full sequence log-likelihoods to capture an entire trajectory of the model's decisions, while methods like GRPO aggregate per-token ratios into a trajectory-level update. However, both often limit exploration on long-horizon tasks. We introduce History-Aggregated Exploratory Policy Optimization (HAEPO), a history-…