Tootfinder

NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scale
NextStep Team, Chunrui Han, Guopeng Li, Jingwei Wu, Quan Sun, Yan Cai, Yuang Peng, Zheng Ge, Deyu Zhou, Haomiao Tang, Hongyu Zhou, Kenkun Liu, Ailin Huang, Bin Wang, Changxin Miao, Deshan Sun, En Yu, Fukun Yin, Gang Yu, Hao Nie, Haoran Lv, Hanpeng Hu, Jia Wang, Jian Zhou, Jianjian Sun, Kaijun Tan, Kang An, Kangheng Lin, Liang Zhao, Mei Chen, Peng Xing, Rui Wang, Shiyu Liu, Shutao Xia, Tianhao You, Wei Ji…

NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scale
Prevailing autoregressive (AR) models for text-to-image generation either rely on heavy, computationally-intensive diffusion models to process continuous image tokens, or employ vector quantization (VQ) to obtain discrete tokens with quantization loss. In this paper, we push the autoregressive paradigm forward with NextStep-1, a 14B autoregressive model paired with a 157M flow matching head, training on discrete text tokens and continuous image tokens with next-token prediction objectives. Next…

WebWeaver: Structuring Web-Scale Evidence with Dynamic Outlines for Open-Ended Deep Research
Zijian Li, Xin Guan, Bo Zhang, Shen Huang, Houquan Zhou, Shaopeng Lai, Ming Yan, Yong Jiang, Pengjun Xie, Fei Huang, Jun Zhang, Jingren Zhou
https://arxiv.org/abs/2509.13312

WebWeaver: Structuring Web-Scale Evidence with Dynamic Outlines for Open-Ended Deep Research
This paper tackles open-ended deep research (OEDR), a complex challenge where AI agents must synthesize vast web-scale information into insightful reports. Current approaches are plagued by dual-fold limitations: static research pipelines that decouple planning from evidence acquisition and one-shot generation paradigms that easily suffer from long-context failure issues like "loss in the middle" and hallucinations. To address these challenges, we introduce WebWeaver, a novel dual-agent framewo…

High-efficiency and long-distance quantum memory-assisted device-independent quantum secret sharing with single photon sources
Qi Zhang, Jia-Wei Ying, Shi-Pu Gu, Xing-Fu Wang, Lan Zhou, Yu-Bo Sheng
https://arxiv.org/abs/2510.12288

High-efficiency and long-distance quantum memory-assisted device-independent quantum secret sharing with single photon sources
Quantum secret sharing (QSS) plays a critical role in building the distributed quantum networks. Device-independent (DI) QSS provides the highest security level for QSS. However, the photon transmission loss and extremely low multipartite entanglement generation rate largely limit DI QSS's secure photon transmission distance and practical key generation efficiency. To address the above drawbacks, we propose the quantum memory-assisted (QMA) DI QSS protocol based on single photon sources (SPSs).…

Self-Supervised Temporal Super-Resolution of Energy Data using Generative Adversarial Transformer
Xuanhao Mu, G\"okhan Demirel, Yuzhe Zhang, Jianlei Liu, Thorsten Schlachter, Veit Hagenmeyer
https://arxiv.org/abs/2508.10587

Self-Supervised Temporal Super-Resolution of Energy Data using Generative Adversarial Transformer
To bridge the temporal granularity gap in energy network design and operation based on Energy System Models, resampling of time series is required. While conventional upsampling methods are computationally efficient, they often result in significant information loss or increased noise. Advanced models such as time series generation models, Super-Resolution models and imputation models show potential, but also face fundamental challenges. The goal of time series generative models is to learn the…

Differentially Private Synthetic Text Generation for Retrieval-Augmented Generation (RAG)
Junki Mori, Kazuya Kakizaki, Taiki Miyagawa, Jun Sakuma
https://arxiv.org/abs/2510.06719

Differentially Private Synthetic Text Generation for Retrieval-Augmented Generation (RAG)
Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by grounding them in external knowledge. However, its application in sensitive domains is limited by privacy risks. Existing private RAG methods typically rely on query-time differential privacy (DP), which requires repeated noise injection and leads to accumulated privacy loss. To address this issue, we propose DP-SynRAG, a framework that uses LLMs to generate differentially private synthetic RAG databases. Unlike prior…

Uplink RSMA for Pinching-Antenna Systems
Apostolos A. Tegos, Yue Xiao, Sotiris A. Tegos, George K. Karagiannidis, Panagiotis D. Diamantoulakis
https://arxiv.org/abs/2509.10076 h…

Uplink RSMA for Pinching-Antenna Systems
One of the key goals of next-generation wireless networks is to adapt to changing conditions and meet the growing demand for reliable, high-capacity communications from emerging applications. Overcoming the limitations of conventional technologies, such as fixed antenna positions, is essential to achieving this objective because it mitigates the impact of path loss on the received signal and creates strong line-of-sight links, enhancing system performance. With this in mind, the newly proposed …

Loss-tolerant parallelized Bell-state generation with a hybrid cat qudit
Z. M. McIntyre, W. A. Coish
https://arxiv.org/abs/2509.08577 https://arxiv.org/pdf…

Loss-tolerant parallelized Bell-state generation with a hybrid cat qudit
Having multiple Bell pairs shared by distant quantum registers provides a key resource for both quantum networks and distributed quantum computing. In this paper, we present a protocol for parallelized Bell-pair generation that uses the phase of a coherent light pulse to encode a qudit, enabling the simultaneous generation of multiple Bell pairs. By encoding a qudit in a basis of light-matter Schrödinger's cat states, the loss of a photon in transit can be detected through an $XX$ parity syndr…

Loss Functions for Detecting Outliers in Panel Data
Charles D. Coleman, Thomas Bryan
https://arxiv.org/abs/2509.07014 https://arxiv.org/pdf/2509.07014

Loss Functions for Detecting Outliers in Panel Data
The detection of outliers is of critical importance in the assurance of data quality. Outliers may exist in observed data or in data derived from these observed data, such as estimates and forecasts. An outlier may indicate a problem with its data generation process or may simply be a true statement about the world. Without making any distributional assumptions, this paper proposes the use of loss functions to detect these outliers in panel data. An unsigned loss function is derived axiomatical…

Squeezed Light Generation in Periodically Poled Thin-Film Lithium Niobate Waveguides
Xiaodong Shi, Angela Anna Baiju, Xu Chen, Sakthi Sanjeev Mohanraj, Sihao Wang, Veerendra Dhyani, Biveen Shajilal, Mengyao Zhao, Ran Yang, Yue Li, Guangxing Wu, Hao Hao, Victor Leong, Ping Koy Lam, Di Zhu
https://arxiv.org/abs/2508.08599

Squeezed Light Generation in Periodically Poled Thin-Film Lithium Niobate Waveguides
Squeezed states of light play a key role in quantum-enhanced sensing and continuous-variable quantum information processing. Realizing integrated squeezed light sources is crucial for developing compact and scalable photonic quantum systems. In this work, we demonstrate on-chip broadband vacuum squeezing at telecommunication wavelengths on the thin-film lithium niobate (TFLN) platform. Our device integrates periodically poled lithium niobate (PPLN) nanophotonic waveguides with low-loss edge cou…

APML: Adaptive Probabilistic Matching Loss for Robust 3D Point Cloud Reconstruction
Sasan Sharifipour, Constantino \'Alvarez Casado, Mohammad Sabokrou, Miguel Bordallo L\'opez
https://arxiv.org/abs/2509.08104

APML: Adaptive Probabilistic Matching Loss for Robust 3D Point Cloud Reconstruction
Training deep learning models for point cloud prediction tasks such as shape completion and generation depends critically on loss functions that measure discrepancies between predicted and ground-truth point sets. Commonly used functions such as Chamfer Distance (CD), HyperCD, and InfoCD rely on nearest-neighbor assignments, which often induce many-to-one correspondences, leading to point congestion in dense regions and poor coverage in sparse regions. These losses also involve non-differentiab…

🐛 Using nanotechnology to target crop-munching pests and spare beneficial bugs
https://phys.org/news/2025-09-nanotechnology-crop-munching-pests-beneficial.html

Using nanotechnology to target crop-munching pests and spare beneficial bugs
A bane of farmers' existence, it's estimated that plant-eating pests are responsible for the loss of up to 40% of pre-harvest yields globally. But a new generation of crop treatments that target only "bad" bugs could mean big gains for the Canadian agriculture sector, improving pest management tools in an industry that in 2024 generated over $142 billion.

Operator-Consistent Physics-Informed Learning for Wafer Thermal Reconstruction in Lithography
Ze Tao, Yuxi Jin, Ke Xu, Haoran Xu, Hanxuan Wang, Fujun Liu
https://arxiv.org/abs/2510.09207

Operator-Consistent Physics-Informed Learning for Wafer Thermal Reconstruction in Lithography
Thermal field reconstruction in post-exposure bake (PEB) is critical for advanced lithography, but current methods lack consistent accuracy and physical alignment. To bridge this gap, we introduce a physics-informed neural model that unifies coordinates, fields, and differential operators on a single computation graph using gated liquid layers. Applied to a 2D PEB scenario with internal heat generation and convective boundaries, the model formulates residuals via differential forms and a loss f…

Nonlinear periodic orbit solutions and their bifurcation structure at the origin of soliton hopping in coupled microresonators
Savyaraj Deshmukh, Aleksandr Tusnin, Alexey Tikan, Tobias J. Kippenberg, Tobias M. Schneider
https://arxiv.org/abs/2508.09921

Nonlinear periodic orbit solutions and their bifurcation structure at the origin of soliton hopping in coupled microresonators
Microresonator frequency combs, essential for future integrated optical systems, rely on dissipative Kerr solitons generated in a single microresonator to achieve coherent frequency comb generation. Recent advances in the nanofabrication of low-loss integrated nonlinear microresonators have paved the way for the exploration of coupled-resonator systems. These systems provide significant technological advantages, including higher conversion efficiency and the generation of dual dispersive waves.…

$\Delta L$ Normalization: Rethink Loss Aggregation in RLVR
Zhiyuan He, Xufang Luo, Yike Zhang, Yuqing Yang, Lili Qiu
https://arxiv.org/abs/2509.07558 https://

$ΔL$ Normalization: Rethink Loss Aggregation in RLVR
We propose $ΔL$ Normalization, a simple yet effective loss aggregation method tailored to the characteristic of dynamic generation lengths in Reinforcement Learning with Verifiable Rewards (RLVR). Recently, RLVR has demonstrated strong potential in improving the reasoning capabilities of large language models (LLMs), but a major challenge lies in the large variability of response lengths during training, which leads to high gradient variance and unstable optimization. Although previous methods…

The loss tolerance of cat breeding for fault-tolerant grid state generation
Olga Solodovnikova, Ulrik L. Andersen, Jonas S. Neergaard-Nielsen
https://arxiv.org/abs/2508.06193 ht…

The loss tolerance of cat breeding for fault-tolerant grid state generation
The development of a continuous-variable photonic quantum computer depends on the reliable preparation of high-quality Gottesman-Kitaev-Preskill states. The most promising GKP preparation scheme is the cat breeding protocol, which can generate GKP states deterministically given a source of squeezed cat states, using beam splitters, homodyne detectors and a feedforward displacement. However, analyzing the performance of the protocol under loss is cumbersome due to the exponential scaling of the …

SALSA-V: Shortcut-Augmented Long-form Synchronized Audio from Videos
Amir Dellali, Luca A. Lanzend\"orfer, Florian Gr\"otschla, Roger Wattenhofer
https://arxiv.org/abs/2510.02916

SALSA-V: Shortcut-Augmented Long-form Synchronized Audio from Videos
We propose SALSA-V, a multimodal video-to-audio generation model capable of synthesizing highly synchronized, high-fidelity long-form audio from silent video content. Our approach introduces a masked diffusion objective, enabling audio-conditioned generation and the seamless synthesis of audio sequences of unconstrained length. Additionally, by integrating a shortcut loss into our training process, we achieve rapid generation of high-quality audio samples in as few as eight sampling steps, pavi…

Second-Generation Planet Formation in Post-AGB Discs: Testing the Role of Gravitational Instability
Ali Pourmand, Devika Kamath, Orsola De Marco, Mark Wardle
https://arxiv.org/abs/2509.03894

Second-Generation Planet Formation in Post-AGB Discs: Testing the Role of Gravitational Instability
Post-Asymptotic Giant Branch (post-AGB) binary stars are evolved systems that host circumbinary discs formed through mass loss during late-stage binary interaction. Their structural, kinematic, and chemical similarities to planet-forming discs have motivated the idea that these systems may act as sites of "second-generation" planet formation. In this study, we assess whether the gravitational instability mechanism - a proposed pathway for giant planet formation in some protoplanetary discs-can …

Lang-PINN: From Language to Physics-Informed Neural Networks via a Multi-Agent Framework
Xin He, Liangliang You, Hongduan Tian, Bo Han, Ivor Tsang, Yew-Soon Ong
https://arxiv.org/abs/2510.05158

Lang-PINN: From Language to Physics-Informed Neural Networks via a Multi-Agent Framework
Physics-informed neural networks (PINNs) provide a powerful approach for solving partial differential equations (PDEs), but constructing a usable PINN remains labor-intensive and error-prone. Scientists must interpret problems as PDE formulations, design architectures and loss functions, and implement stable training pipelines. Existing large language model (LLM) based approaches address isolated steps such as code generation or architecture suggestion, but typically assume a formal PDE is alre…

Language Generation in the Limit: Noise, Loss, and Feedback
Yannan Bai, Debmalya Panigrahi, Ian Zhang
https://arxiv.org/abs/2507.15319 https://

Language Generation in the Limit: Noise, Loss, and Feedback
Kleinberg and Mullainathan (2024) recently proposed a formal framework called language generation in the limit and showed that given a sequence of example strings from an unknown target language drawn from any countable collection, an algorithm can correctly generate unseen strings from the target language within finite time. This notion was further refined by Li, Raman, and Tewari (2024), who defined stricter categories of non-uniform and uniform generation. They showed that a finite union of …

Retrieval Feedback Memory Enhancement Large Model Retrieval Generation Method
Leqian Li, Dianxi Shi, Jialu Zhou, Xinyu Wei, Mingyue Yang, Songchang Jin, Shaowu Yang
https://arxiv.org/abs/2508.17862

Retrieval Feedback Memory Enhancement Large Model Retrieval Generation Method
Large Language Models (LLMs) have shown remarkable capabilities across diverse tasks, yet they face inherent limitations such as constrained parametric knowledge and high retraining costs. Retrieval-Augmented Generation (RAG) augments the generation process by retrieving externally stored knowledge absent from the models internal parameters. However, RAG methods face challenges such as information loss and redundant retrievals during multi-round queries, accompanying the difficulties in precise…

DRACO: Data Replication and Collection Framework for Enhanced Data Availability and Robustness in IoT Networks
Waleed Bin Qaim, Oznur Ozkasap, Rabia Qadar, Moncef Gabbouj
https://arxiv.org/abs/2510.07464

DRACO: Data Replication and Collection Framework for Enhanced Data Availability and Robustness in IoT Networks
The Internet of Things (IoT) bridges the gap between the physical and digital worlds, enabling seamless interaction with real-world objects via the Internet. However, IoT systems face significant challenges in ensuring efficient data generation, collection, and management, particularly due to the resource-constrained and unreliable nature of connected devices, which can lead to data loss. This paper presents DRACO (Data Replication and Collection), a framework that integrates a distributed hop-…

Data-Driven Loss Functions for Inference-Time Optimization in Text-to-Image Generation
Sapir Esther Yiflach, Yuval Atzmon, Gal Chechik
https://arxiv.org/abs/2509.02295 https://

Data-Driven Loss Functions for Inference-Time Optimization in Text-to-Image Generation
Text-to-image diffusion models can generate stunning visuals, yet they often fail at tasks children find trivial--like placing a dog to the right of a teddy bear rather than to the left. When combinations get more unusual--a giraffe above an airplane--these failures become even more pronounced. Existing methods attempt to fix these spatial reasoning failures through model fine-tuning or test-time optimization with handcrafted losses that are suboptimal. Rather than imposing our assumptions abou…

mmKey: Channel-Aware Beam Shaping for Reliable Key Generation in mmWave Wireless Networks
Poorya Mollahosseini, Yasaman Ghasempour
https://arxiv.org/abs/2508.19010 https://

mmKey: Channel-Aware Beam Shaping for Reliable Key Generation in mmWave Wireless Networks
Physical-layer key generation (PLKG) has emerged as a promising technique to secure next-generation wireless networks by exploiting the inherent properties of the wireless channel. However, PLKG faces fundamental challenges in the millimeter wave (mmWave) regime due to channel sparsity, higher phase noise, and higher path loss, which undermine both the randomness and reciprocity required for secure key generation. In this paper, we present mmKey, a novel PLKG framework that capitalizes on the a…

Electrical control of quantum dots in GaAs-on-insulator waveguides for coherent single-photon generation
Hanna Salamon, Ying Wang, Arnulf Snedker-Nielsen, Atefeh Shadmani, R\"udiger Schott, Mircea Balauroiu, Nicolas Volet, Arne Ludwig, Leonardo Midolo
https://arxiv.org/abs/2508.04584

Electrical control of quantum dots in GaAs-on-insulator waveguides for coherent single-photon generation
The integration of coherent quantum emitters with silicon photonic platforms essential for scalable quantum technologies. We demonstrate electrically controlled self-assembled quantum dots embedded in GaAs waveguides bonded onto a SiO2/Si substrate and coupled to low-loss SiN waveguides. Our approach uses a die-to-die adhesive bonding process to realize a GaAs-on-insulator platform incorporating a p-i-n junction for charge noise suppression and Stark tuning of excitonic transitions. Resonance f…

One of our final topics in the #ISE2025 lecture were Knowledge Graph Embeddings. How to vectorise KG structures while preserving their inherent semantics?
#AI #KGE

"X of Information'' Continuum: A Survey on AI-Driven Multi-dimensional Metrics for Next-Generation Networked Systems
Beining Wu, Jun Huang, Shui Yu
https://arxiv.org/abs/2507.19657

"X of Information'' Continuum: A Survey on AI-Driven Multi-dimensional Metrics for Next-Generation Networked Systems
The development of next-generation networking systems has inherently shifted from throughput-based paradigms towards intelligent, information-aware designs that emphasize the quality, relevance, and utility of transmitted information, rather than sheer data volume. While classical network metrics, such as latency and packet loss, remain significant, they are insufficient to quantify the nuanced information quality requirements of modern intelligent applications, including autonomous vehicles, d…

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…

Optimizing Speech Language Models for Acoustic Consistency
Morteza Rohanian, Michael Krauthammer
https://arxiv.org/abs/2509.26276 https://arxiv.org/pdf/250…

Optimizing Speech Language Models for Acoustic Consistency
We study speech language models that incorporate semantic initialization and planning losses to achieve robust and consistent generation. Our approach initializes speech tokens with self-supervised features, applies a light alignment loss, and trains with thinning and auxiliary objectives that target robustness and content planning. We train three models: a 0.7B speech-only model, a 1.0B speech-only model, and a 1.0B interleaved model with both text and speech. Acoustic studies show that the sp…

CharGen: Fast and Fluent Portrait Modification
Jan-Niklas Dihlmann, Arnela Killguss, Hendrik P. A. Lensch
https://arxiv.org/abs/2509.25058 https://arxiv.or…

CharGen: Fast and Fluent Portrait Modification
Interactive editing of character images with diffusion models remains challenging due to the inherent trade-off between fine-grained control, generation speed, and visual fidelity. We introduce CharGen, a character-focused editor that combines attribute-specific Concept Sliders, trained to isolate and manipulate attributes such as facial feature size, expression, and decoration with the StreamDiffusion sampling pipeline for more interactive performance. To counteract the loss of detail that oft…

Active RISs: Modeling and Optimization
Recep Akif Tasci, Panagiotis Gavriilidis, Ertugrul Basar, George C. Alexandropoulos
https://arxiv.org/abs/2507.16499 https://

Active RISs: Modeling and Optimization
Reconfigurable Intelligent Surfaces (RIS)-empowered communication has emerged as a transformative technology for next generation wireless networks, enabling the programmable shaping of the propagation environment. However, conventional RISs are fundamentally limited by the double path loss effect, which severely attenuates the reflected signals. To overcome this, active RIS architectures, capable of amplifying impinging signals, have been proposed. This chapter investigates the modeling, perfor…

Frequency comb generation in low-loss, low-stress, high-Q deuterated silicon nitride microring resonators in an 8-inch photonics platform
Y. Cao, G. F. Chen, C. Lau, L. Y. M. Tobing, S. L. H. Jang, Y. F. Tsang, J. O. Yoo, Y. T. Toh, J. S. Goh, L. W. Lim, C. W. Wong, D. K. T. Ng, D. T. H. Tan, X. Luo
https://arxiv.org/abs/2507.17287

Frequency comb generation in low-loss, low-stress, high-Q deuterated silicon nitride microring resonators in an 8-inch photonics platform
Systematic studies on different SiN films in terms of propagation losses are presented, and deuterated SiN emerges as a good candidate for ultralow loss (< 0.1 dB/cm) and reliability by simple 8-inch process with low thermal budget. Frequency comb generation in high-Q (~1 million) deuterated silicon nitride microring is demonstrated and used for intensity modulated direct detection transmission. Negligible power penalty for 25.78 GBaud/s NRZ and PAM4 is achieved at error rates <10-6, below the …

Dispatch-Aware Deep Neural Network for Optimal Transmission Switching: Toward Real-Time and Feasibility Guaranteed Operation
Minsoo Kim, Jip Kim
https://arxiv.org/abs/2507.17194

Dispatch-Aware Deep Neural Network for Optimal Transmission Switching: Toward Real-Time and Feasibility Guaranteed Operation
Optimal transmission switching (OTS) improves optimal power flow (OPF) by selectively opening transmission lines, but its mixed-integer formulation increases computational complexity, especially on large grids. To deal with this, we propose a dispatch-aware deep neural network (DA-DNN) that accelerates DC-OTS without relying on pre-solved labels. DA-DNN predicts line states and passes them through a differentiable DC-OPF layer, using the resulting generation cost as the loss function so that al…

Optimizing What Matters: AUC-Driven Learning for Robust Neural Retrieval
Nima Sheikholeslami, Erfan Hosseini, Patrice Bechard, Srivatsava Daruru, Sai Rajeswar
https://arxiv.org/abs/2510.00137

Optimizing What Matters: AUC-Driven Learning for Robust Neural Retrieval
Dual-encoder retrievers depend on the principle that relevant documents should score higher than irrelevant ones for a given query. Yet the dominant Noise Contrastive Estimation (NCE) objective, which underpins Contrastive Loss, optimizes a softened ranking surrogate that we rigorously prove is fundamentally oblivious to score separation quality and unrelated to AUC. This mismatch leads to poor calibration and suboptimal performance in downstream tasks like retrieval-augmented generation (RAG).…

Sea-ing Through Scattered Rays: Revisiting the Image Formation Model for Realistic Underwater Image Generation
Vasiliki Ismiroglou, Malte Pedersen, Stefan H. Bengtson, Andreas Aakerberg, Thomas B. Moeslund
https://arxiv.org/abs/2509.15011

Sea-ing Through Scattered Rays: Revisiting the Image Formation Model for Realistic Underwater Image Generation
In recent years, the underwater image formation model has found extensive use in the generation of synthetic underwater data. Although many approaches focus on scenes primarily affected by discoloration, they often overlook the model's ability to capture the complex, distance-dependent visibility loss present in highly turbid environments. In this work, we propose an improved synthetic data generation pipeline that includes the commonly omitted forward scattering term, while also considering a …

Imbalance in Balance: Online Concept Balancing in Generation Models
Yukai Shi, Jiarong Ou, Rui Chen, Haotian Yang, Jiahao Wang, Xin Tao, Pengfei Wan, Di Zhang, Kun Gai
https://arxiv.org/abs/2507.13345

Imbalance in Balance: Online Concept Balancing in Generation Models
In visual generation tasks, the responses and combinations of complex concepts often lack stability and are error-prone, which remains an under-explored area. In this paper, we attempt to explore the causal factors for poor concept responses through elaborately designed experiments. We also design a concept-wise equalization loss function (IMBA loss) to address this issue. Our proposed method is online, eliminating the need for offline dataset processing, and requires minimal code changes. In o…

Mind the Gap: Conformative Decoding to Improve Output Diversity of Instruction-Tuned Large Language Models
Max Peeperkorn, Tom Kouwenhoven, Dan Brown, Anna Jordanous
https://arxiv.org/abs/2507.20956

Mind the Gap: Conformative Decoding to Improve Output Diversity of Instruction-Tuned Large Language Models
Instruction-tuning large language models (LLMs) reduces the diversity of their outputs, which has implications for many tasks, particularly for creative tasks. This paper investigates the ``diversity gap'' for a writing prompt narrative generation task. This gap emerges as measured by current diversity metrics for various open-weight and open-source LLMs. The results show significant decreases in diversity due to instruction-tuning. We explore the diversity loss at each fine-tuning stage for th…

Robust quantum communication through lossy microwave links
James D. Teoh, Nathanael Cottet, Patrick Winkel, Luke D. Burkhart, Luigi Frunzio, Robert J. Schoelkopf
https://arxiv.org/abs/2509.18547

1-GHz VIS-to-MIR frequency combs enabled by CMOS-compatible nanophotonic waveguides
Xuan Zhang, Yuchen Wang, Junguo Xu, Guangzi Feng, Xia Hou, Qiankun Li, Xueying Sun, Yongyuan Chu, Chengbo Mou, Hairun Guo, Sida Xing
https://arxiv.org/abs/2509.23704

1-GHz VIS-to-MIR frequency combs enabled by CMOS-compatible nanophotonic waveguides
A fully stabilized frequency comb is essential for precision metrology and coherent optical synthesis. However, fully-stabilized frequency combs generally require separate stages for supercontinuum generation (SCG) and self-referencing, largely limiting their compactness. Here, enabled by the low-threshold multi-octave supercontinuum generation and concurrent third-harmonic generation in low-loss silicon nitride waveguides, we present a novel approach to a self-referenced frequency comb source …

Spatially tailored spin wave excitation for spurious-free, low-loss magnetostatic wave filters with ultra-wide frequency tunability
Shuxian Wu, Shun Yao, Xingyu Du, Chin-Yu Chang, Roy H. Olsson III
https://arxiv.org/abs/2507.14469

Spatially tailored spin wave excitation for spurious-free, low-loss magnetostatic wave filters with ultra-wide frequency tunability
Yttrium iron garnet magnetostatic wave (MSW) radio frequency (RF) cavity filters are promising for sixth-generation (6G) communication systems due to their wide frequency tunability. However, the presence of severe spurious modes arising from the finite cavity dimensions severely degrades the filter performance. We present a half-cone transducer that spatially tailors spin wave excitation to selectively enhance the primary cavity modes comprising the MSW filter passband, while strongly suppress…

Approaching transform-limited linewidths in telecom-wavelength transitions of ungated quantum dots
Andrew N. Wakileh, Dan Dalacu, Philip J. Poole, Boris Lamontagne, Simona Moisa, Robin L. Williams, Nir Rotenberg
https://arxiv.org/abs/2509.02320

Approaching transform-limited linewidths in telecom-wavelength transitions of ungated quantum dots
Highly coherent quantum emitters operating in the telecommunication C-band (1530 - 1565nm), where ultra-low-loss fibers and photonic circuits are available, are crucial to the development of scalable quantum technologies. In this work, we report on a modified Stranski-Krastanov growth scheme using chemical beam epitaxy to enable the generation of high-quality InAs/InP quantum dots, characterized by near-transform-limited linewidths ($Γ_{\mathrm{TL}}$). We demonstrate the growth of highly-symme…

Eyes on the Image: Gaze Supervised Multimodal Learning for Chest X-ray Diagnosis and Report Generation
Tanjim Islam Riju, Shuchismita Anwar, Saman Sarker Joy, Farig Sadeque, Swakkhar Shatabda
https://arxiv.org/abs/2508.13068

Eyes on the Image: Gaze Supervised Multimodal Learning for Chest X-ray Diagnosis and Report Generation
We propose a two-stage multimodal framework that enhances disease classification and region-aware radiology report generation from chest X-rays, leveraging the MIMIC-Eye dataset. In the first stage, we introduce a gaze-guided contrastive learning architecture for disease classification. It integrates visual features, clinical labels, bounding boxes, and radiologist eye-tracking signals and is equipped with a novel multi-term gaze-attention loss combining MSE, KL divergence, correlation, and cen…

Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Models
Donghoon Kim, Dongyoung Lee, Ik Joon Chang, Sung-Ho Bae
https://arxiv.org/abs/2509.26436

Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Models
Diffusion models achieve high-quality image generation but face deployment challenges due to their high computational requirements. Although 8-bit outlier-aware post-training quantization (PTQ) matches full-precision performance, extending PTQ to 4 bits remains challenging. Larger step sizes in 4-bit quantization amplify rounding errors in dense, low-magnitude activations, leading to the loss of fine-grained textures. We hypothesize that not only outliers but also small activations are critical…

Physics-Informed Diffusion Models for Unsupervised Anomaly Detection in Multivariate Time Series
Juhi Soni, Markus Lange-Hegermann, Stefan Windmann
https://arxiv.org/abs/2508.11528

Physics-Informed Diffusion Models for Unsupervised Anomaly Detection in Multivariate Time Series
We propose an unsupervised anomaly detection approach based on a physics-informed diffusion model for multivariate time series data. Over the past years, diffusion model has demonstrated its effectiveness in forecasting, imputation, generation, and anomaly detection in the time series domain. In this paper, we present a new approach for learning the physics-dependent temporal distribution of multivariate time series data using a weighted physics-informed loss during diffusion model training. A …

Early Stopping Chain-of-thoughts in Large Language Models
Minjia Mao, Bowen Yin, Yu Zhu, Xiao Fang
https://arxiv.org/abs/2509.14004 https://arxiv.org/pdf/2…

Early Stopping Chain-of-thoughts in Large Language Models
Reasoning large language models (LLMs) have demonstrated superior capacities in solving complicated problems by generating long chain-of-thoughts (CoT), but such a lengthy CoT incurs high inference costs. In this study, we introduce ES-CoT, an inference-time method that shortens CoT generation by detecting answer convergence and stopping early with minimal performance loss. At the end of each reasoning step, we prompt the LLM to output its current final answer, denoted as a step answer. We then…

On the Trade-Off Between Sum-Rate and Energy Efficiency through the Convergence of HAPS and Active RIS Technologies
Bilal Karaman, Ilhan Basturk, Ferdi Kara, Metin Ozturk, Sezai Taskin, Halil Yanikomeroglu
https://arxiv.org/abs/2507.13889

On the Trade-Off Between Sum-Rate and Energy Efficiency through the Convergence of HAPS and Active RIS Technologies
This paper investigates the integration of active reconfigurable intelligent surfaces (RIS) relay with high-altitude platform stations (HAPS) to enhance non-terrestrial network (NTN) performance in next-generation wireless systems. While prior studies focused on passive RIS architectures, the severe path loss and double fading in long-distance HAPS links make active RIS a more suitable alternative due to its inherent signal amplification capabilities. We formulate a sum-rate maximization proble…

Accurate and Scalable Simulation of Cavity-Based Networks in Modular Quantum Architectures
Sahar Ben Rached, Zezhou Sun, Guilu Long, Santiago Rodrigo, Carmen G. Almud\'ever, Eduard Alarc\'on, Sergi Abadal
https://arxiv.org/abs/2508.13896

Accurate and Scalable Simulation of Cavity-Based Networks in Modular Quantum Architectures
Cavity-mediated interconnects are a promising platform for scaling modular quantum computers by enabling high-fidelity inter-chip quantum state transmission and entanglement generation. In this work, we first model the dynamics of deterministic inter-chip quantum state transfer using the Stimulated Raman Adiabatic Passage (STIRAP) protocol, analyzing fidelity loss mechanisms under experimentally achievable qubit-cavity coupling and decoherence parameters. We then extend the NetSquid simulator, …