Tootfinder

This https://arxiv.org/abs/2310.02003 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csSE_…

L2MAC: Large Language Model Automatic Computer for Extensive Code Generation
Transformer-based large language models (LLMs) are constrained by the fixed context window of the underlying transformer architecture, hindering their ability to produce long and coherent outputs. Memory-augmented LLMs are a promising solution, but current approaches cannot handle long output generation tasks since they (1) only focus on reading memory and reduce its evolution to the concatenation of new memories or (2) use very specialized memories that cannot adapt to other domains. This pape…

ChuXin: 1.6B Technical Report
Xiaomin Zhuang, Yufan Jiang, Qiaozhi He, Zhihua Wu
https://arxiv.org/abs/2405.04828 https://arxiv.org/p…

ChuXin: 1.6B Technical Report
In this report, we present ChuXin, an entirely open-source language model with a size of 1.6 billion parameters. Unlike the majority of works that only open-sourced the model weights and architecture, we have made everything needed to train a model available, including the training data, the training process, and the evaluation code. Our goal is to empower and strengthen the open research community, fostering transparency and enabling a new wave of innovation in the field of language modeling. …

GCV-Turbo: End-to-end Acceleration of GNN-based Computer Vision Tasks on FPGA
Bingyi Zhang, Rajgopal Kannan, Carl Busart, Viktor Prasanna
https://arxiv.org/abs/2404.07188

GCV-Turbo: End-to-end Acceleration of GNN-based Computer Vision Tasks on FPGA
Graph neural networks (GNNs) have recently empowered various novel computer vision (CV) tasks. In GNN-based CV tasks, a combination of CNN layers and GNN layers or only GNN layers are employed. This paper introduces GCV-Turbo, a domain-specific accelerator on FPGA for end-to-end acceleration of GNN-based CV tasks. GCV-Turbo consists of two key components: (1) a \emph{novel} hardware architecture optimized for the computation kernels in both CNNs and GNNs using the same set of computation resour…

Set-Encoder: Permutation-Invariant Inter-Passage Attention for Listwise Passage Re-Ranking with Cross-Encoders
Ferdinand Schlatt, Maik Fr\"obe, Harrisen Scells, Shengyao Zhuang, Bevan Koopman, Guido Zuccon, Benno Stein, Martin Potthast, Matthias Hagen
https://arxiv.org/abs/2404.06912

Set-Encoder: Permutation-Invariant Inter-Passage Attention for Listwise Passage Re-Ranking with Cross-Encoders
Cross-encoders are effective passage re-rankers. But when re-rank\-ing multiple passages at once, existing cross-encoders inefficiently optimize the output ranking over several input permutations, as their passage interactions are not permutation-invariant. Moreover, their high memory footprint constrains the number of passages during listwise training. To tackle these issues, we propose the Set-Encoder, a new cross-encoder architecture that (1) introduces inter-passage attention with parallel …

ChuXin: 1.6B Technical Report
Xiaomin Zhuang, Yufan Jiang, Qiaozhi He, Zhihua Wu
https://arxiv.org/abs/2405.04828 https://arxiv.org/p…

ChuXin: 1.6B Technical Report
In this report, we present ChuXin, an entirely open-source language model with a size of 1.6 billion parameters. Unlike the majority of works that only open-sourced the model weights and architecture, we have made everything needed to train a model available, including the training data, the training process, and the evaluation code. Our goal is to empower and strengthen the open research community, fostering transparency and enabling a new wave of innovation in the field of language modeling. …

GCV-Turbo: End-to-end Acceleration of GNN-based Computer Vision Tasks on FPGA
Bingyi Zhang, Rajgopal Kannan, Carl Busart, Viktor Prasanna
https://arxiv.org/abs/2404.07188

GCV-Turbo: End-to-end Acceleration of GNN-based Computer Vision Tasks on FPGA
Graph neural networks (GNNs) have recently empowered various novel computer vision (CV) tasks. In GNN-based CV tasks, a combination of CNN layers and GNN layers or only GNN layers are employed. This paper introduces GCV-Turbo, a domain-specific accelerator on FPGA for end-to-end acceleration of GNN-based CV tasks. GCV-Turbo consists of two key components: (1) a \emph{novel} hardware architecture optimized for the computation kernels in both CNNs and GNNs using the same set of computation resour…

Set-Encoder: Permutation-Invariant Inter-Passage Attention for Listwise Passage Re-Ranking with Cross-Encoders
Ferdinand Schlatt, Maik Fr\"obe, Harrisen Scells, Shengyao Zhuang, Bevan Koopman, Guido Zuccon, Benno Stein, Martin Potthast, Matthias Hagen
https://arxiv.org/abs/2404.06912

Set-Encoder: Permutation-Invariant Inter-Passage Attention for Listwise Passage Re-Ranking with Cross-Encoders
Cross-encoders are effective passage re-rankers. But when re-rank\-ing multiple passages at once, existing cross-encoders inefficiently optimize the output ranking over several input permutations, as their passage interactions are not permutation-invariant. Moreover, their high memory footprint constrains the number of passages during listwise training. To tackle these issues, we propose the Set-Encoder, a new cross-encoder architecture that (1) introduces inter-passage attention with parallel …

Digital models and sensory substitution devices: An inclusive communication of architecture through sound https://link.springer.com/chapter/10.1007/978-3-031-57575-4_40

Digital Models and Sensory Substitution Devices: An Inclusive Communication of Architecture Through Sound
The singular nature of architecture, compared to other types of artistic experience, admits the spatial component as the main reference for understanding an object. While scientific advances in the production of digital models have allowed for more and more...

How a new SF permit could enable more hostile architecture. https://www.axios.com/local/san-francisco/2024/02/22/sf-hostile-architecture-permit-love-our-neighbors
It’s being enabled already since there’s no enforcement of the deca…

How a new SF permit could enable more "hostile architecture"
The permit, aimed at enhancing neighborhoods, could lead to more planters as some locals move to deter homeless encampments.

xAI open sources the base model weights and network architecture of Grok-1, a 314B parameter Mixture-of-Experts model, under the Apache 2.0 license (x.ai)
https://x.ai/blog/grok-os

Mars 2.0: A Toolchain for Modeling, Analysis, Verification and Code Generation of Cyber-Physical Systems
Bohua Zhan, Xiong Xu, Qiang Gao, Zekun Ji, Xiangyu Jin, Shuling Wang, Naijun Zhan
https://arxiv.org/abs/2403.03035

Mars 2.0: A Toolchain for Modeling, Analysis, Verification and Code Generation of Cyber-Physical Systems
We introduce Mars 2.0 for modeling, analysis, verification and code generation of Cyber-Physical Systems. Mars 2.0 integrates Mars 1.0 with several important extensions and improvements, allowing the design of cyber-physical systems using the combination of AADL and Simulink/Stateflow, which provide a unified graphical framework for modeling the functionality, physicality and architecture of the system to be developed. For a safety-critical system, formal analysis and verification of its combin…

TeraPool-SDR: An 1.89TOPS 1024 RV-Cores 4MiB Shared-L1 Cluster for Next-Generation Open-Source Software-Defined Radios
Yichao Zhang, Marco Bertuletti, Samuel Riedel, Matheus Cavalcante, Alessandro Vanelli-Coralli, Luca Benini
https://arxiv.org/abs/2405.04988

TeraPool-SDR: An 1.89TOPS 1024 RV-Cores 4MiB Shared-L1 Cluster for Next-Generation Open-Source Software-Defined Radios
Radio Access Networks (RAN) workloads are rapidly scaling up in data processing intensity and throughput as the 5G (and beyond) standards grow in number of antennas and sub-carriers. Offering flexible Processing Elements (PEs), efficient memory access, and a productive parallel programming model, many-core clusters are a well-matched architecture for next-generation software-defined RANs, but staggering performance requirements demand a high number of PEs coupled with extreme Power, Performance…

Scattering of Giant Planets and Implications for the Origin of the Hierarchical and Eccentric Two-planet System GJ 1148
Longhui Yuan (HKU, ETH Zurich), Man Hoi Lee (HKU)
https://arxiv.org/abs/2405.04045

Scattering of Giant Planets and Implications for the Origin of the Hierarchical and Eccentric Two-planet System GJ 1148
The GJ 1148 system has two Saturn-mass planets orbiting around an M dwarf star on hierarchical and eccentric orbits, with orbital period ratio of 13 and eccentricities of both planets of 0.375. The inner planet is in the regime of eccentric warm Jupiters. We perform numerical experiments to study the planet-planet scattering scenario for the origin of this orbital architecture. We consider a third planet of $0.1 M_J$ (Jupiter's mass) in the initial GJ 1148 system with initial orbital separation…

This https://arxiv.org/abs/2402.01116 has been replaced.
link: https://scholar.google.com/scholar?q=a

Scalable Multi-modal Model Predictive Control via Duality-based Interaction Predictions
We propose a hierarchical architecture designed for scalable real-time Model Predictive Control (MPC) in complex, multi-modal traffic scenarios. This architecture comprises two key components: 1) RAID-Net, a novel attention-based Recurrent Neural Network that predicts relevant interactions along the MPC prediction horizon between the autonomous vehicle and the surrounding vehicles using Lagrangian duality, and 2) a reduced Stochastic MPC problem that eliminates irrelevant collision avoidance co…

March 17, 2024
Open Release of Grok-1
We are releasing the weights and architecture of our 314 billion parameter Mixture-of-Experts model, Grok-1.
#musk

Grok is here and it's opensource (the model), curious how bad it is , trained on X and other data 😃
#ai

This https://arxiv.org/abs/2402.01831 has been replaced.
link: https://scholar.google.com/scholar?q=a

Audio Flamingo: A Novel Audio Language Model with Few-Shot Learning and Dialogue Abilities
Augmenting large language models (LLMs) to understand audio -- including non-speech sounds and non-verbal speech -- is critically important for diverse real-world applications of LLMs. In this paper, we propose Audio Flamingo, a novel audio language model with 1) strong audio understanding abilities, 2) the ability to quickly adapt to unseen tasks via in-context learning and retrieval, and 3) strong multi-turn dialogue abilities. We introduce a series of training techniques, architecture design…

This https://arxiv.org/abs/1809.02467 has been replaced.
link: https://scholar.google.com/scholar?q=a

Design and testing of 3D-printed micro-architectured polymer materials exhibiting a negative Poisson's ratio
This work proposes the complete design cycle for several auxetic materials where the cycle consists of three steps (i) the design of the micro-architecture, (ii) the manufacturing of the material and (iii) the testing of the material. We use topology optimization via a level-set method and asymptotic homogenization to obtain periodic micro-architectured materials with a prescribed effective elasticity tensor and Poisson's ratio. The space of admissible micro-architectural shapes that carries or…

This https://arxiv.org/abs/2312.17279 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…

Stateful Conformer with Cache-based Inference for Streaming Automatic Speech Recognition
In this paper, we propose an efficient and accurate streaming speech recognition model based on the FastConformer architecture. We adapted the FastConformer architecture for streaming applications through: (1) constraining both the look-ahead and past contexts in the encoder, and (2) introducing an activation caching mechanism to enable the non-autoregressive encoder to operate autoregressively during inference. The proposed model is thoughtfully designed in a way to eliminate the accuracy disp…

Faster Convergence for Transformer Fine-tuning with Line Search Methods
Philip Kenneweg, Leonardo Galli, Tristan Kenneweg, Barbara Hammer
https://arxiv.org/abs/2403.18506

Faster Convergence for Transformer Fine-tuning with Line Search Methods
Recent works have shown that line search methods greatly increase performance of traditional stochastic gradient descent methods on a variety of datasets and architectures [1], [2]. In this work we succeed in extending line search methods to the novel and highly popular Transformer architecture and dataset domains in natural language processing. More specifically, we combine the Armijo line search with the Adam optimizer and extend it by subdividing the networks architecture into sensible units…

Neural network reconstruction of density and velocity fields from the 2MASS Redshift Survey
Robert Lilow, Punyakoti Ganeshaiah Veena, Adi Nusser
https://arxiv.org/abs/2404.02278 <…

Neural network reconstruction of density and velocity fields from the 2MASS Redshift Survey
We reconstruct the 3D matter density and peculiar velocity fields in the local Universe up to a distance of $200\,h^{-1}\,\mathrm{Mpc}$ from the Two-Micron All-Sky Redshift Survey (2MRS), using a neural network (NN). We employ a NN with U-net autoencoder architecture and a weighted mean squared error loss function, trained separately to output either the density or velocity field for a given input grid of galaxy number counts. The NN is trained on mocks derived from the Quijote N-body simulatio…

Pseudorandom Permutations from Random Reversible Circuits
William He, Ryan O'Donnell
https://arxiv.org/abs/2404.14648 https://arx…

Pseudorandom Permutations from Random Reversible Circuits
We study pseudorandomness properties of permutations on $\{0,1\}^n$ computed by random circuits made from reversible $3$-bit gates (permutations on $\{0,1\}^3$). Our main result is that a random circuit of depth $n \cdot \tilde{O}(k^2)$, with each layer consisting of $\approx n/3$ random gates in a fixed nearest-neighbor architecture, yields almost $k$-wise independent permutations. The main technical component is showing that the Markov chain on $k$-tuples of $n$-bit strings induced by a singl…

Extended Galerkin neural network approximation of singular variational problems with error control
Mark Ainsworth, Justin Dong
https://arxiv.org/abs/2405.00815

Extended Galerkin neural network approximation of singular variational problems with error control
We present extended Galerkin neural networks (xGNN), a variational framework for approximating general boundary value problems (BVPs) with error control. The main contributions of this work are (1) a rigorous theory guiding the construction of new weighted least squares variational formulations suitable for use in neural network approximation of general BVPs (2) an ``extended'' feedforward network architecture which incorporates and is even capable of learning singular solution structures, thus…

Semantic Information Theory in a feedback-control Kuramoto Model
Damian R Sowinski, Adam Frank, Gourab Ghoshal
https://arxiv.org/abs/2404.02221 https://

Semantic Information Theory in a feedback-control Kuramoto Model
Semantic Information Theory (SIT) offers a new approach to evaluating the information architecture of complex systems. In this study we describe the steps required to {\it operationalize} SIT via its application to dynamical problems. Our road map has four steps: (1) separating the dynamical system into agent-environment sub-systems; (2) choosing an appropriate coarse graining and quantifying correlations; (3) identifying a measure of viability; (4) implementing a scrambling protocol and measur…

This https://arxiv.org/abs/2404.10404 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCR_…

Sisu: Decentralized Trustless Bridge For Full Ethereum Node
In this paper, we present a detailed approach and implementation to prove Ethereum full node using recursive SNARK, distributed general GKR and Groth16. Our protocol's name is Sisu whose architecture is based on distributed Virgo in zkBridge with some major improvements. Besides proving signature aggregation, we provide solutions to 2 hard problems in proving Ethereum full node: 1) any public key is valid under previous beacon state and 2) all public keys are pairwise distinct. Our solution doe…

This https://arxiv.org/abs/2401.10417 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csAR_…

SSR: Spatial Sequential Hybrid Architecture for Latency Throughput Tradeoff in Transformer Acceleration
With the increase in the computation intensity of the chip, the mismatch between computation layer shapes and the available computation resource significantly limits the utilization of the chip. Driven by this observation, prior works discuss spatial accelerators or dataflow architecture to maximize the throughput. However, using spatial accelerators could potentially increase the execution latency. In this work, we first systematically investigate two execution models: (1) sequentially (tempor…

hartverscheurend, die prachtige afbeeldingen van Rotterdam uit begin vorige eeuw van @…
https://mastodon.ozioso.online/@Photoc

This https://arxiv.org/abs/2402.01116 has been replaced.
link: https://scholar.google.com/scholar?q=a

Scalable Multi-modal Model Predictive Control via Duality-based Interaction Predictions
We propose a hierarchical architecture designed for scalable real-time Model Predictive Control (MPC) in complex, multi-modal traffic scenarios. This architecture comprises two key components: 1) RAID-Net, a novel attention-based Recurrent Neural Network that predicts relevant interactions along the MPC prediction horizon between the autonomous vehicle and the surrounding vehicles using Lagrangian duality, and 2) a reduced Stochastic MPC problem that eliminates irrelevant collision avoidance co…

This https://arxiv.org/abs/2403.15965 has been replaced.
initial toot: https://mastoxiv.page/@arX…

Controllable bipolaron formation unveiling structural features of trap states in organic charge transport
Magnetic resonance methods offer a unique chance for in-depth study of conductive organic material systems, not only accounts for number of charge carriers, but also allows manipulations of spin dynamics of particles. Here we present a study of continuous-wave electrically detected magnetic resonance on a range of organic conjugate polymers under transistor architecture, with tunability in both carrier concentration and drifting electric field. We demonstrate the general existence of bipolaron …

Accelerating Innovation in 6G Research: Real-Time Capable SDR System Architecture for Rapid Prototyping
Maximilian Engelhardt, Sebastian Giehl, Carsten Andrich, Michael Schubert, Christian Schneider, Alexander Ebert, Markus Landmann, Giovanni Del Galdo
https://arxiv.org/abs/2402.06520

Accelerating Innovation in 6G Research: Real-Time Capable SDR System Architecture for Rapid Prototyping
The next global mobile communication standard 6G strives to push the technological limits of radio frequency (RF) communication even further than its predecessors: Data rates beyond 100 Gbit/s, RF bandwidths above 1 GHz, and sub-millisecond latency necessitate very high performance development tools to enable the extent of innovation required for 6G's likely features. We propose a new SDR firmware and software architecture designed explicitly to meet these challenging requirements. It relies on…

This https://arxiv.org/abs/2402.05226 has been replaced.
link: https://scholar.google.com/scholar?q=a

A Comprehensive Analysis of Secondary Coexistence in a Real-World CBRS Deployment
The Federal Communications Commission (FCC) in the U.S. has made the Citizens Broadband Radio Service (CBRS) band (3.55 - 3.7 GHz) available for commercial wireless usage under a shared approach using a three-tier hierarchical architecture, where the federal incumbent is the highest priority Tier 1 user, Priority Access License (PAL) holders, who have paid for licenses, are Tier 2 users and Tier 3 users operate under General Authorized Access (GAA), without license fees or protection from highe…

Hamiltonian engineering with time-ordered evolution for unitary control of electron spins in semiconductor quantum dots
Bohdan Khromets, Zach D. Merino, Jonathan Baugh
https://arxiv.org/abs/2402.08146

Hamiltonian engineering with time-ordered evolution for unitary control of electron spins in semiconductor quantum dots
We present a unitary control pulse design method for a scalable quantum computer architecture based on electron spins in lateral quantum dots. We employ simultaneous control of spin interactions and derive the functional forms of spin Hamiltonian parameter pulses for a universal set of 1- and 2-qubit logic gates. This includes selective spin rotations with the weak local g-factor variations in the presence of the global oscillating field, and a Control-Phase operation with the simultaneous cont…

This https://arxiv.org/abs/2310.02003 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csSE_…

L2MAC: Large Language Model Automatic Computer for Extensive Code Generation
Transformer-based large language models (LLMs) are constrained by the fixed context window of the underlying transformer architecture, hindering their ability to produce long and coherent outputs. Memory-augmented LLMs are a promising solution, but current approaches cannot handle long output generation tasks since they (1) only focus on reading memory and reduce its evolution to the concatenation of new memories or (2) use very specialized memories that cannot adapt to other domains. This pape…

A Modified T-Mixer Simulation, Fabrication, and Characterization For 1D Diffusion Controlled Studies
M. Rashed Khan, Taliman Afroz, Caleb Shaw
https://arxiv.org/abs/2402.14998

A Modified T-Mixer Simulation, Fabrication, and Characterization For 1D Diffusion Controlled Studies
Here, we describe a simple and unique architecture of a microfluidic mixer that can mix two streams (Water and Fluorescein Isothiocyanate (FITC) buffer) with 30% efficiency. The overall mixing in this design is dominated by 1D diffusion, and to enhance mixing, we used three different types of geometric obstacles inside the channel. Comsol multiphysics simulation software was used to validate the theoretical mixing efficiency (at Reynolds Number, Re ~ 0.1, 1, and 10) of this device. We utilized …

This https://arxiv.org/abs/2404.10404 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCR_…

Sisu: Decentralized Trustless Bridge For Full Ethereum Node
In this paper, we present a detailed approach and implementation to prove Ethereum full node using recursive SNARK, distributed general GKR and Groth16. Our protocol's name is Sisu whose architecture is based on distributed Virgo in zkBridge with some major improvements. Besides proving signature aggregation, we provide solutions to 2 hard problems in proving Ethereum full node: 1) any public key is valid under previous beacon state and 2) all public keys are pairwise distinct. Our solution doe…

Highly stable photoluminescence in vacuum-processed halide perovskite core-shell 1D nanostructures
Javier Castillo-Seoane, Lidia Contreras-Bernal, T. Cristina Rojas, Juan P. Espinos, Andres-Felipe Castro-Mendez, Juan-P. Correa-Baena, Angel Barranco, Juan R. Sanchez-Valencia, Ana Borras
https://arxiv.org/abs/2402.19269

Highly stable photoluminescence in vacuum-processed halide perovskite core-shell 1D nanostructures
Hybrid organometal halide perovskites (HP) present exceptional optoelectronic properties, but their poor long-term stability is a major bottleneck for their commercialization. Herein, we present a solvent-free approach to growing single-crystal organic nanowires (ONW), nanoporous metal oxide scaffolds, and HP to form a core@multishell architecture. The synthetic procedure is carried out under mild vacuum conditions employing thermal evaporation for the metal-free phthalocyanine (H2Pc) nanowires…

1-bit Quantized On-chip Hybrid Diffraction Neural Network Enabled by Authentic All-optical Fully-connected Architecture
Yu Shao, Haiqi Gao, Yipeng Chen, Yujie liu, Junren Wen, Haidong He, Yuchuan Shao, Yueguang Zhang, Weidong Shen, Chenying Yang
https://arxiv.org/abs/2404.07443

1-bit Quantized On-chip Hybrid Diffraction Neural Network Enabled by Authentic All-optical Fully-connected Architecture
Optical Diffraction Neural Networks (DNNs), a subset of Optical Neural Networks (ONNs), show promise in mirroring the prowess of electronic networks. This study introduces the Hybrid Diffraction Neural Network (HDNN), a novel architecture that incorporates matrix multiplication into DNNs, synergizing the benefits of conventional ONNs with those of DNNs to surmount the modulation limitations inherent in optical diffraction neural networks. Utilizing a singular phase modulation layer and an ampli…

This https://arxiv.org/abs/2402.05226 has been replaced.
link: https://scholar.google.com/scholar?q=a

A Comprehensive Analysis of Secondary Coexistence in a Real-World CBRS Deployment
The Federal Communications Commission (FCC) in the U.S. has made the Citizens Broadband Radio Service (CBRS) band (3.55 - 3.7 GHz) available for commercial wireless usage under a shared approach using a three-tier hierarchical architecture, where the federal incumbent is the highest priority Tier 1 user, Priority Access License (PAL) holders, who have paid for licenses, are Tier 2 users and Tier 3 users operate under General Authorized Access (GAA), without license fees or protection from highe…

This https://arxiv.org/abs/2401.15379 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_ees…

Time-modulated phased array controlled with non-ideal bipolar squared periodic sequences
Bipolar (+/-1) sequences with no zero state suit particularly well for safeguarding the switched feeding network efficiency when applied to time-modulated arrays (TMAs). During the zero state of a conventional time-modulating sequence, if a given array element is switched off, a certain amount of energy of the transmitted/received signal is wasted. We propose a novel single sideband time-modulated phased array (TMPA) architecture governed by realistic bipolar squared sequences in which the rise…

Sisu: Decentralized Trustless Bridge For Full Ethereum Node
Billy Pham, Huy Nguyen
https://arxiv.org/abs/2404.10404 https://arxiv.org…

Sisu: Decentralized Trustless Bridge For Full Ethereum Node
In this paper, we present a detailed approach and implementation to prove Ethereum full node using recursive SNARK, distributed general GKR and Groth16. Our protocol's name is Sisu whose architecture is based on distributed Virgo in zkBridge with some major improvements. Besides proving signature aggregation, we provide solutions to 2 hard problems in proving Ethereum full node: 1) any public key is valid under previous beacon state and 2) all public keys are pairwise distinct. Our solution doe…

This https://arxiv.org/abs/2404.09841 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_ees…

Anatomy of Industrial Scale Multilingual ASR
This paper describes AssemblyAI's industrial-scale automatic speech recognition (ASR) system, designed to meet the requirements of large-scale, multilingual ASR serving various application needs. Our system leverages a diverse training dataset comprising unsupervised (12.5M hours), supervised (188k hours), and pseudo-labeled (1.6M hours) data across four languages. We provide a detailed description of our model architecture, consisting of a full-context 600M-parameter Conformer encoder pre-trai…

One-step architecture of bifunctional petal-like oxygen-deficient NiAl-LDHs nanosheets for high-performance hybrid supercapacitors and urea oxidation
Yuchen Wang, Yaoyu Liu, Man Zhang, Biying Liu, Zhiyue Zhao, Kai Yan
https://arxiv.org/abs/2403.17038

One-step architecture of bifunctional petal-like oxygen-deficient NiAl-LDHs nanosheets for high-performance hybrid supercapacitors and urea oxidation
Nickel-based layered double hydroxides (LDHs) are promising electrode materials in the fields of energy storage (supercapacitors) and conversion (urea oxidation). The rational construction of atomic and electronic structure is crucial for nickel-based LDHs to realize their satisfactory electrochemical performance. Herein, we report a facile, ecofriendly, one-step synthesis process to construct petal-like oxygen-deficient NiAl-LDH nanosheets for hybrid super-capacitors (HSCs) and urea oxidation …

This https://arxiv.org/abs/2403.09539 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…

Logits of API-Protected LLMs Leak Proprietary Information
The commercialization of large language models (LLMs) has led to the common practice of high-level API-only access to proprietary models. In this work, we show that even with a conservative assumption about the model architecture, it is possible to learn a surprisingly large amount of non-public information about an API-protected LLM from a relatively small number of API queries (e.g., costing under $1,000 for OpenAI's gpt-3.5-turbo). Our findings are centered on one key observation: most moder…

This https://arxiv.org/abs/2307.16089 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csIR_…

Train Once, Use Flexibly: A Modular Framework for Multi-Aspect Neural News Recommendation
Recent neural news recommenders (NNRs) extend content-based recommendation (1) by aligning additional aspects (e.g., topic, sentiment) between candidate news and user history or (2) by diversifying recommendations w.r.t. these aspects. This customization is achieved by "hardcoding" additional constraints into the NNR's architecture and/or training objectives: any change in the desired recommendation behavior thus requires retraining the model with a modified objective. This impedes widespread a…

This https://arxiv.org/abs/2403.09539 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…

Logits of API-Protected LLMs Leak Proprietary Information
The commercialization of large language models (LLMs) has led to the common practice of high-level API-only access to proprietary models. In this work, we show that even with a conservative assumption about the model architecture, it is possible to learn a surprisingly large amount of non-public information about an API-protected LLM from a relatively small number of API queries (e.g., costing under $1,000 for OpenAI's gpt-3.5-turbo). Our findings are centered on one key observation: most moder…

Controllable bipolaron formation unveiling structural features of trap states in organic charge transport
Zichen Wang, Ilia Kulikov, Tarig Mustafa, Jan Behrends, Henning Sirringhaus
https://arxiv.org/abs/2403.15965

Controllable bipolaron formation unveiling structural features of trap states in organic charge transport
Magnetic resonance methods offer a unique chance for in-depth study of conductive organic material systems, not only accounts for number of charge carriers, but also allows manipulations of spin dynamics of particles. Here we present a study of continuous-wave electrically detected magnetic resonance on a range of organic conjugate polymers under transistor architecture, with tunability in both carrier concentration and drifting electric field. We demonstrate the general existence of bipolaron …

This https://arxiv.org/abs/2404.05892 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCL_…

Eagle and Finch: RWKV with Matrix-Valued States and Dynamic Recurrence
We present Eagle (RWKV-5) and Finch (RWKV-6), sequence models improving upon the RWKV (RWKV-4) architecture. Our architectural design advancements include multi-headed matrix-valued states and a dynamic recurrence mechanism that improve expressivity while maintaining the inference efficiency characteristics of RNNs. We introduce a new multilingual corpus with 1.12 trillion tokens and a fast tokenizer based on greedy matching for enhanced multilinguality. We trained four Eagle models, ranging fr…

This https://arxiv.org/abs/2404.06912 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csIR_…

Set-Encoder: Permutation-Invariant Inter-Passage Attention for Listwise Passage Re-Ranking with Cross-Encoders
Cross-encoders are effective passage re-rankers. But when re-rank\-ing multiple passages at once, existing cross-encoders inefficiently optimize the output ranking over several input permutations, as their passage interactions are not permutation-invariant. Moreover, their high memory footprint constrains the number of passages during listwise training. To tackle these issues, we propose the Set-Encoder, a new cross-encoder architecture that (1) introduces inter-passage attention with parallel …

This https://arxiv.org/abs/2404.02221 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_nli…

Semantic Information Theory in a feedback-control Kuramoto Model
Semantic Information Theory (SIT) offers a new approach to evaluating the information architecture of complex systems. In this study we describe the steps required to {\it operationalize} SIT via its application to dynamical problems. Our road map has four steps: (1) separating the dynamical system into agent-environment sub-systems; (2) choosing an appropriate coarse graining and quantifying correlations; (3) identifying a measure of viability; (4) implementing a scrambling protocol and measur…

Optomechanically Induced Transparency/Absorption in a 3D Microwave Cavity Architecture at Ambient Temperature
Sumit Kumar, Matthew Kenworthy, Henry Ginn, Xavier Rojas
https://arxiv.org/abs/2402.10935

Optomechanically Induced Transparency/Absorption in a 3D Microwave Cavity Architecture at Ambient Temperature
Leveraging advancements in cavity optomechanics, we explore Optomechanically Induced Transparency/Absorption (OMIT/OMIA) in the microwave domain at ambient temperature. Contrary to previous works employing cryogenic temperatures, this work exploits a 3D microwave cavity architecture to observe these effects at ambient temperature, broadening the scope of possible applications. The work successfully enhances the optomechanical coupling strength, enabling observable and robust OMIT/OMIA effects, …

Free-Floating Planets, Survivor Planets, Captured Planets and Binary Planets from Stellar Flybys
Fangyuan Yu, Dong Lai
https://arxiv.org/abs/2403.07224 htt…

Free-Floating Planets, Survivor Planets, Captured Planets and Binary Planets from Stellar Flybys
In star clusters, close stellar encounters can strongly impact the architecture of a planetary system or even destroy it. We present a systematic study on the effects of stellar flybys on two-planet systems. When such a system (with a modest initial planet semi-major axis ratio, $a_1/a_2=0.6-0.8$) experiences stellar flybys, one or both planets can be ejected, forming free-floating planets (FFPs), captured planets (CPs) around the flyby star, and free-floating binary planets (BPs); the remainin…

Performance analysis of photodetectors based on 2D materials and heterostructures
Nagaraj Nandihalli
https://arxiv.org/abs/2402.11375 https://

Performance analysis of photodetectors based on 2D materials and heterostructures
The unprecedented demand for sophisticated, self-powered, compact, ultrafast, cost-effective, and broadband light sensors for a myriad of applications has spurred a lot of research, precipitating in a slew of studies over the last decade. Apart from the photosensing ability of an active element in the light sensor, the device architecture is crucial in terms of photoinduced charge carrier generation and separation. Since the inception of graphene and the subsequent research growth in the atomic…

This https://arxiv.org/abs/2402.19269 has been replaced.
initial toot: https://mastoxiv.page/@arX…

Highly stable photoluminescence in vacuum-processed halide perovskite core-shell 1D nanostructures
Hybrid organometal halide perovskites (HP) present exceptional optoelectronic properties, but their poor long-term stability is a major bottleneck for their commercialization. Herein, we present a solvent-free approach to growing single-crystal organic nanowires (ONW), nanoporous metal oxide scaffolds, and HP to form a core@multishell architecture. The synthetic procedure is carried out under mild vacuum conditions employing thermal evaporation for the metal-free phthalocyanine (H2Pc) nanowires…