prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
tensor network methods for quantum-inspired image processing and
classical
quantum dot
tensor network methods for quantum-inspired image processing and classical optics
evaluation with a human-aligned gpt-judge and a user study with 108 students shows that 8b-scribe models achieve comparable or superior quality to much
larger
smaller models
evaluation with a human-aligned gpt-judge and a user study with 108 students shows that 8b-scribe models achieve comparable or superior quality to much larger models in key dimensions such as relevance and actionability while being perceived on par with gpt-4o and llama-3
through the judge s eyes inferred thinking
traces
llm raters
through the judge s eyes inferred thinking traces improve reliability of llm raters
we further extend these analyses to two broad families of activation functions and deep feedforward architectures demonstrating that
abstract
neural networks
we further extend these analyses to two broad families of activation functions and deep feedforward architectures demonstrating that abstract representations naturally arise in all these scenarios
in this paper we propose csi2q a novel csi fingerprinting system that
achieves
csi dataset
in this paper we propose csi2q a novel csi fingerprinting system that achieves comparable performance to iq-based approaches
the effectiveness of the approach is demonstrated through
simulations
multi-drone racing
the effectiveness of the approach is demonstrated through simulations and real-world experiments with the crazyflie quadrotor
reinforcement learning for pollution detection in a randomized sparse and
nonstationary
reinforcement learning rl
reinforcement learning for pollution detection in a randomized sparse and nonstationary environment with an autonomous underwater vehicle
our work incorporates state-of-the-art innovations in all key aspects of the fault-tolerant quantum
computing
quantum coherence
our work incorporates state-of-the-art innovations in all key aspects of the fault-tolerant quantum computing ftqc stack to show how quantum computers could realistically and practically tackle co _2 utilization for green energy production
we solve this by treating multinomial updates as independent square-root diffusions of zero drift yielding a closed-form
law
stochastic differential
we solve this by treating multinomial updates as independent square-root diffusions of zero drift yielding a closed-form law for the first-extinction time
neural-network-based controllers nncs can represent complex highly nonlinear control laws but verifying the closed-loop
stability
data-driven stabilization
neural-network-based controllers nncs can represent complex highly nonlinear control laws but verifying the closed-loop stability of dynamical systems using them remains challenging
our findings reveal that both the fermi level and the bi-isotropic response offer independent and complementary control parameters for tuning the strength of light--magnon coupling in ti afm heterostructures with potential implications for reconfigurable thz spintronic and
photonic
photonic circuits
our findings reveal that both the fermi level and the bi-isotropic response offer independent and complementary control parameters for tuning the strength of light--magnon coupling in ti afm heterostructures with potential implications for reconfigurable thz spintronic and photonic devices
cross-lingual alignment cla aims to align
multilingual
multilingual data
cross-lingual alignment cla aims to align multilingual representations enabling large language models llms to seamlessly transfer knowledge across languages
in this work we report momentum-resolved reflectivity measurements on
photonic
photonic devices
in this work we report momentum-resolved reflectivity measurements on photonic crystals that are periodic in two dimensions and homogeneous over a thickness of 5 mu m
the tri-infrastructure methodology and 79-problem benchmark enable longitudinal tracking of
reasoning
reasoning curriculum
the tri-infrastructure methodology and 79-problem benchmark enable longitudinal tracking of reasoning capabilities as foundation models evolve
using a simulation model based on the prisoner s dilemma and quantal response equilibrium we analyze how variations in these
learning
network structures
using a simulation model based on the prisoner s dilemma and quantal response equilibrium we analyze how variations in these learning rates affect the emergence of large-scale network structures
to address this we propose a dual-conditioning framework that combines eeg embeddings with spatial saliency maps to
enhance
image generation
to address this we propose a dual-conditioning framework that combines eeg embeddings with spatial saliency maps to enhance image generation
to solve these issues we propose a novel hybrid model of bph and he
distributions
distribution shift
to solve these issues we propose a novel hybrid model of bph and he distributions borrowing the most desirable features from each for enhanced approximation quality
galaxy mergers trigger starburst activity and galactic outflows that enrich the
circumgalactic
galaxy cgm
galaxy mergers trigger starburst activity and galactic outflows that enrich the circumgalactic medium profoundly impacting galaxy evolution
using a simulation model based on the prisoner s dilemma and quantal response equilibrium we analyze how variations in these
learning
scale-free networks
using a simulation model based on the prisoner s dilemma and quantal response equilibrium we analyze how variations in these learning rates affect the emergence of large-scale network structures
by quantifying relevance through uncertainty the approach enables generalizable inference across tasks including generative inference
causal
causal inference
by quantifying relevance through uncertainty the approach enables generalizable inference across tasks including generative inference causal discovery anomaly detection and time series forecasting
by rejecting foundationalist quests for a single essential definition of personhood this paper offers a more pragmatic and flexible way to think about integrating
ai
artificial intelligence
by rejecting foundationalist quests for a single essential definition of personhood this paper offers a more pragmatic and flexible way to think about integrating ai agents into our society
interestingly for the trapped system a cat-like superposition state corresponds to maximum entanglement
entropy
multipartite entanglement
interestingly for the trapped system a cat-like superposition state corresponds to maximum entanglement entropy below the transition highlighting the relevance of the present model for studying the effect of decoherence on intra-particle entanglement in the context of quantum information processing
this manifests in three critical challenges 1 inefficient unguided exploration 2 imprecise credit assignment due to overlooking pivotal states and 3 myopic planning caused by static
reward
reward density
this manifests in three critical challenges 1 inefficient unguided exploration 2 imprecise credit assignment due to overlooking pivotal states and 3 myopic planning caused by static reward discounting
a tidal disruption event tde occurs when a star passes within the tidal radius of a
supermassive
black hole
a tidal disruption event tde occurs when a star passes within the tidal radius of a supermassive black hole smbh
we introduce see4d a pose-free trajectory-to-camera framework that replaces explicit trajectory prediction with rendering to a bank of fixed virtual cameras thereby separating
camera
pose estimation
we introduce see4d a pose-free trajectory-to-camera framework that replaces explicit trajectory prediction with rendering to a bank of fixed virtual cameras thereby separating camera control from scene modeling
reality distortion room rdr is a proof-of-concept augmented
reality
virtual reality
reality distortion room rdr is a proof-of-concept augmented reality system using projection mapping and unencumbered interaction with the microsoft roomalive system to study a user s locomotive response to visual effects that seemingly transform the physical room the user is in
experiments demonstrate that asyncthink achieves 28 lower inference latency compared to parallel thinking while improving accuracy on
mathematical
reasoning tasks
experiments demonstrate that asyncthink achieves 28 lower inference latency compared to parallel thinking while improving accuracy on mathematical reasoning
we extend these local convergence results to locally uniformly
convex
strongly convex
we extend these local convergence results to locally uniformly convex functions and fully adaptive methods which do not need knowledge of the lipschitz constant thus providing the first sharp local rates for ar p
recent advances in visual question answering vqa have demonstrated impressive performance in natural image domains with models like llava leveraging large
language
vision-language models
recent advances in visual question answering vqa have demonstrated impressive performance in natural image domains with models like llava leveraging large language models llms for open-ended reasoning
this features combined together prevent the use of the classical mild solution theory of
hjb
hjb equation
this features combined together prevent the use of the classical mild solution theory of hjb equation see e
a famous conjecture of goemans on single-source unsplittable flows states that one can turn any fractional
flow
maximum flow
a famous conjecture of goemans on single-source unsplittable flows states that one can turn any fractional flow into an unsplittable one of no higher cost while increasing the load on any arc by at most the maximum demand
identifying and addressing security issues during the
early
security issues
identifying and addressing security issues during the early phase of the development lifecycle is critical for mitigating the long-term negative impacts on software systems
we aim at exploring the existence of a stellar mass-metallicity relation of bulges mz r and analyze the possible imprint of characteristics features by accretion and migration of
stars
bulge stars
we aim at exploring the existence of a stellar mass-metallicity relation of bulges mz r and analyze the possible imprint of characteristics features by accretion and migration of stars which could store information on their assembly histories
for each input ccmp adaptively combines class-specific
prompts
prompt tuning
for each input ccmp adaptively combines class-specific prompts using weights derived from global class prototypes and client class priors
keywords small language models factual grounding directed
reasoning
natural language
keywords small language models factual grounding directed reasoning fine-tuning model alignment cost-efficient ai
the recently proposed core-kg framework addresses these limitations by integrating a type-aware
coreference
coreference resolution
the recently proposed core-kg framework addresses these limitations by integrating a type-aware coreference module and domain-guided structured prompts significantly reducing node duplication and legal noise
qualitative changes in spontaneous emission decay as system
parameters
spontaneous emission
qualitative changes in spontaneous emission decay as system parameters are varied
the cross dual encoder network comprises four essential components a global encoder a local
encoder
dual encoder
the cross dual encoder network comprises four essential components a global encoder a local encoder a symmetric cross-attention module and a flow-based decoder
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained bilevel
optimization
bilevel optimization
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained bilevel optimization sflcb
specifically to assess how well llms reason with normative modals we make a comparison between their reasoning with normative modals and their
reasoning
reasoning capabilities
specifically to assess how well llms reason with normative modals we make a comparison between their reasoning with normative modals and their reasoning with epistemic modals which share a common formal structure
stacked intelligent surfaces sis are a promising technology for next-generation
wireless
reconfigurable intelligent surface
stacked intelligent surfaces sis are a promising technology for next-generation wireless systems offering an opportunity to enhance communication performance with low power consumption
in the circumgalactic medium cgm and the intracluster medium icm it is expected that relativistic cosmic
rays
circumgalactic medium
in the circumgalactic medium cgm and the intracluster medium icm it is expected that relativistic cosmic rays collide with thermal particles and produce gamma -rays through pion decay
large language models llms show strong potential to support creative
tasks
large language
large language models llms show strong potential to support creative tasks but the role of the interface design is poorly understood
through hierarchical goal alignment task assignment and conflict resolution orchvis enables
humans
ai agents
through hierarchical goal alignment task assignment and conflict resolution orchvis enables humans to supervise complex multi-agent workflows without micromanaging each step
the function f is assumed to be em r -decomposable meaning there exist m ge1 subsets v_1 dots v_m of v each with a cardinality at most r and a corresponding set of nonnegative supermodular
functions
monotone submodular
the function f is assumed to be em r -decomposable meaning there exist m ge1 subsets v_1 dots v_m of v each with a cardinality at most r and a corresponding set of nonnegative supermodular functions f_i 2 v_i rightarrow mathbb r _ i 1 ldots m such that f s sum_ i 1 m f_i s cap v_i holds for each s subseteq v
to formalize this we extend the concept of data informativity by requiring the existence of a controller that stabilizes all systems consistent with the data and the
prior
prior knowledge
to formalize this we extend the concept of data informativity by requiring the existence of a controller that stabilizes all systems consistent with the data and the prior knowledge
- for edge orientation problems we give a partial characterization of the problems that have a randomized complexity of omega log n and the problems that have a
randomized
randomized algorithm
- for edge orientation problems we give a partial characterization of the problems that have a randomized complexity of omega log n and the problems that have a randomized complexity of poly log log n
our work not only reveals the key role of anisotropic epc in controlling the thermal and optical properties of tairte4 but also provides insights into designing polarization-sensitive optoelectronic
devices
photonic devices
our work not only reveals the key role of anisotropic epc in controlling the thermal and optical properties of tairte4 but also provides insights into designing polarization-sensitive optoelectronic devices based on topological semimetals
however significant limitations remain mllms generalize poorly across digital-physical spaces and embodiments vision-language-action models vlas produce low-level actions yet lack robust high-level embodied reasoning and most embodied large
language
vision-language models
however significant limitations remain mllms generalize poorly across digital-physical spaces and embodiments vision-language-action models vlas produce low-level actions yet lack robust high-level embodied reasoning and most embodied large language models ellms are constrained to digital-space with poor generalization...
the dataset is carefully constructed with temporal splits comprehensive
features
real-world datasets
the dataset is carefully constructed with temporal splits comprehensive features and strict leakage prevention to support realistic and reproducible machine learning evaluation
to facilitate valid uncertainty quantification and hypothesis testing on matching decisions we further develop a general debiasing and projection framework for arbitrary linear forms of the
reward
reward models
to facilitate valid uncertainty quantification and hypothesis testing on matching decisions we further develop a general debiasing and projection framework for arbitrary linear forms of the reward matrix deriving asymptotic normality with finite-sample guarantees under matching-induced dependent sampling
this note introduces a unified theory for
causal
causal inference
this note introduces a unified theory for causal inference that integrates riesz regression covariate balancing density-ratio estimation dre targeted maximum likelihood estimation tmle and the matching estimator in average treatment effect ate estimation
communication impact is quantied by a capacity-motivated lower bound obtained from the linear minimum mean-squared error error covariance with a mismatched
channel
communication systems
communication impact is quantied by a capacity-motivated lower bound obtained from the linear minimum mean-squared error error covariance with a mismatched channel estimate
the present study provides a pathway to enhance the ionic conductivity as well as understanding of microscopic conduction mechanism ionic conduction pathways and the role of crystal structure on the
ionic
ionic conduction
the present study provides a pathway to enhance the ionic conductivity as well as understanding of microscopic conduction mechanism ionic conduction pathways and the role of crystal structure on the ionic conduction
our core innovations include 1 vision-language decoupling that moves conventional early vision and
language
vision-language-action vla
our core innovations include 1 vision-language decoupling that moves conventional early vision and language inputs fusion in vlm to late stage achieving better performance while enabling caching and reduce inference overhead and latency 2 long-short action chunking to ensure smooth coherent multi-step planning without ...
drawing on recent research into the neural basis of visual awareness we propose that spatial simulation and perceptual experience depend on shared representational geometries captured by
higher-order
higher-order visual
drawing on recent research into the neural basis of visual awareness we propose that spatial simulation and perceptual experience depend on shared representational geometries captured by higher-order indices of perceptual relations
our calculations show that the resulting kerr
rotation
kerr rotation
our calculations show that the resulting kerr rotation is due to the dc electric field modification of the optical conductivity within a couple of nanometers from the surface consistent with the dependence on the local mirror symmetry breaking at the surface
in contrast imitation learning il is easy to train but often underperforms due to its
offline
imitation learning
in contrast imitation learning il is easy to train but often underperforms due to its offline nature
these tasks are carried out using neural networks and the edge device
seeks
neural network
these tasks are carried out using neural networks and the edge device seeks to optimize inference performance under energy and delay constraints
we ask the following fundamental question is there any application for which a
randomized
randomized algorithm
we ask the following fundamental question is there any application for which a randomized algorithm outperforms any deterministic rom algorithm
however when deployed in real-world scenarios such solutions typically face scalability issues and have to address practical requirements such as full
access
real-world scenarios
however when deployed in real-world scenarios such solutions typically face scalability issues and have to address practical requirements such as full access to original datasets and model
this corresponds to adaptive behavior in the population or through
governmental
adaptive immune
this corresponds to adaptive behavior in the population or through governmental non-pharmaceutical interventions
to investigate the neural representations that emerge in these networks we develop an analytical framework that
maps
receptive fields
to investigate the neural representations that emerge in these networks we develop an analytical framework that maps the optimization over the network weights into a mean-field problem over the distribution of neural preactivations
we show that if the system is controllable then incorporating this as prior knowledge does not relax the
conditions
linear control
we show that if the system is controllable then incorporating this as prior knowledge does not relax the conditions required for data-driven stabilization
this framework can be applied to policy evaluation using the
panel
panel data
this framework can be applied to policy evaluation using the panel data approach pda where we further establish inference for the average treatment effect
the key innovation lies in decoupling the joint optimization into two interleaved phases first updating 3d gaussian parameters via differentiable rendering with fixed poses and second refining camera poses using a customized 3d optical
flow
image fusion
the key innovation lies in decoupling the joint optimization into two interleaved phases first updating 3d gaussian parameters via differentiable rendering with fixed poses and second refining camera poses using a customized 3d optical flow algorithm that incorporates geometric and photometric constraints
we find that despite surfacing errors different
language
language models
we find that despite surfacing errors different language models learn interchangeable representations of numbers that are systematic highly accurate and universal across their hidden states and the types of input contexts
we consider theoretically the possibility of coexisting ferroelectric and metallic altermagnetic order which has recently been predicted in insulating and semiconducting systems via
ab
ab initio calculations
we consider theoretically the possibility of coexisting ferroelectric and metallic altermagnetic order which has recently been predicted in insulating and semiconducting systems via ab initio calculations
the end of manual decoding towards truly end-to-end
language
large language models llms
the end of manual decoding towards truly end-to-end language models
decomposing prediction uncertainty into its aleatoric irreducible and epistemic reducible components is critical for the development and deployment of
machine
debiased machine learning
decomposing prediction uncertainty into its aleatoric irreducible and epistemic reducible components is critical for the development and deployment of machine learning systems
tunable colloidal synthesis enabling μ-arpes on
individual
colloidal synthesis
tunable colloidal synthesis enabling μ-arpes on individual two-dimensional bismuth nanocrystals
for this recurrence time as well as for measures of clonal
diversity
phylogenetic diversity
for this recurrence time as well as for measures of clonal diversity and the size of the largest resistant clone at recurrence we derive corresponding law of large number limits
finally our model shows that cost reduction methods are less effective at reducing
environmental
environmental change
finally our model shows that cost reduction methods are less effective at reducing environmental impact in prejudiced populations
we propose an affective epidemiology framework showing that
collective
opinion dynamics
we propose an affective epidemiology framework showing that collective emotions are governed by network position and volatility rather than personality stability -- transforming how we understand emotional leadership in human systems
conventional single-bus designs inevitably link the conditions for critical coupling a transmission zero and maximum intra-cavity power preventing independent control of these phenomena and restricting the ability to engineer
coupling
coupling regimes
conventional single-bus designs inevitably link the conditions for critical coupling a transmission zero and maximum intra-cavity power preventing independent control of these phenomena and restricting the ability to engineer coupling regimes and resonance lineshapes
designing efficient quantum circuits is a central bottleneck to exploring the potential of quantum
computing
quantum channels
designing efficient quantum circuits is a central bottleneck to exploring the potential of quantum computing particularly for noisy intermediate-scale quantum nisq devices where circuit efficiency and resilience to errors are paramount
our key idea is to decouple visual and linguistic adaptation by introducing two lightweight modules a domain classifier to identify the input image type and a dual adapter mechanism comprising a prompt adapter for language modulation and a visual
adapter
computer vision
our key idea is to decouple visual and linguistic adaptation by introducing two lightweight modules a domain classifier to identify the input image type and a dual adapter mechanism comprising a prompt adapter for language modulation and a visual adapter for vision feature adjustment
hybrid dqn-td3 reinforcement learning for autonomous
navigation
obstacle avoidance
hybrid dqn-td3 reinforcement learning for autonomous navigation in dynamic environments
in decode we grow galaxies with their sfr linked to halo accretion rate distributions via
abundance
stellar mass
in decode we grow galaxies with their sfr linked to halo accretion rate distributions via abundance matching
while the parallel trends assumption is counterfactual and cannot be tested directly researchers often examine pre-treatment periods to check whether the time trends are
parallel
treatment effect boundaries
while the parallel trends assumption is counterfactual and cannot be tested directly researchers often examine pre-treatment periods to check whether the time trends are parallel before treatment is administered
by introducing the twin concepts of reliability and precision along with the corresponding measures mainen and sejnowski s seminal 1995 paper reliability of spike timing in neocortical neurons mainen and sejnowski 1995 paved the way for a new kind of quantitative
spike
spike train
by introducing the twin concepts of reliability and precision along with the corresponding measures mainen and sejnowski s seminal 1995 paper reliability of spike timing in neocortical neurons mainen and sejnowski 1995 paved the way for a new kind of quantitative spike train analysis
we introduce neural stochastic flows nsfs and their latent variants which directly learn latent sde transition laws using conditional normalising
flows
gradient flow
we introduce neural stochastic flows nsfs and their latent variants which directly learn latent sde transition laws using conditional normalising flows with architectural constraints that preserve properties inherited from stochastic flows
free fall diffusion explosion and causal manual
actions
temporal understanding
free fall diffusion explosion and causal manual actions division addition that humans recognize almost instantly
yet a location with lower habitat quality may play a major role in a
species
ecological interactions
yet a location with lower habitat quality may play a major role in a species spread if it acts as a bridge between regions that would otherwise be physically fragmented
in addition the same physical process generates two independent random sequences with identical
entropy
entanglement entropy
in addition the same physical process generates two independent random sequences with identical entropy a private sequence used for secure applications and a public one enabling external statistical verification with zero mutual information between them
we propose a bayesian variable selection method in the framework of modal regression for
heavy-tailed
variable selection
we propose a bayesian variable selection method in the framework of modal regression for heavy-tailed responses
we study the fundamental question of how efficiently suffix array entries can be accessed when the
array
compressed indexing
we study the fundamental question of how efficiently suffix array entries can be accessed when the array cannot be stored explicitly
multi-dataset joint pre-training of emotional
eeg
electroencephalography eeg
multi-dataset joint pre-training of emotional eeg enables generalizable affective computing
through extensive experiments on eight benchmarks we demonstrate that autodeco not only significantly outperforms default
decoding
sparse autoencoders
through extensive experiments on eight benchmarks we demonstrate that autodeco not only significantly outperforms default decoding strategies but also achieves performance comparable to an oracle-tuned baseline derived from hacking the test set -a practical upper bound for any static method
to this end we compare different representation extraction strategies and introduce two model-agnostic
embedding
language models
to this end we compare different representation extraction strategies and introduce two model-agnostic embedding augmentations
we conjecture that similar behavior holds for other structured models including planted
problems
regular graphs
we conjecture that similar behavior holds for other structured models including planted problems in graphs
however in realistic materials where disorder scattering also contributes to nonlinear
transport
transport properties
however in realistic materials where disorder scattering also contributes to nonlinear transport identifying the geometric mechanisms remains a challenge
a limiting case of this scheme utilizing two pulses with identical gaussians envelopes and tuned delay and relative phase is also explored revealing experimentally accessible pathways for manipulating
quantum
quantum technologies
a limiting case of this scheme utilizing two pulses with identical gaussians envelopes and tuned delay and relative phase is also explored revealing experimentally accessible pathways for manipulating quantum coherence
this effect exhibits a lower junction capacitance as compared to its injection counterparts leading to a higher electro-optic bandwidth although the effective
refractive
optical properties
this effect exhibits a lower junction capacitance as compared to its injection counterparts leading to a higher electro-optic bandwidth although the effective refractive index change is low
the fused multimodal representation is input to a logistic regression model which is both
interpretable
predictive processing
the fused multimodal representation is input to a logistic regression model which is both interpretable and computationally efficient
we develop an online algorithm and prove that it achieves a competitive ratio of max 2 min gamma 3 where gamma is the max min storage cost
ratio
competitive ratio
we develop an online algorithm and prove that it achieves a competitive ratio of max 2 min gamma 3 where gamma is the max min storage cost ratio among all servers
the influence of the moving mass is described in body-frame and included as states in both an additional kinematic equation and as part of the coupled rigid-body kinetics of the
underwater
underwater vehicles
the influence of the moving mass is described in body-frame and included as states in both an additional kinematic equation and as part of the coupled rigid-body kinetics of the underwater vehicle
by experimental results pvmark efficiently enables public verifiability on the state-of-the-art llm watermarking schemes yet without compromising the
watermarking
watermark detection
by experimental results pvmark efficiently enables public verifiability on the state-of-the-art llm watermarking schemes yet without compromising the watermarking performance promising to be deployed in practice
neural networks nns offer fast surrogates yet their black-box behavior raises concerns about constraint
violations
soft constraints
neural networks nns offer fast surrogates yet their black-box behavior raises concerns about constraint violations that can compromise safety
taking the u-net connectivity as a base of our framework we evaluate and compare several approaches to improve the segmentation model s architecture and training pipeline including pre-processing techniques encoder backbone
deep
deep learning
taking the u-net connectivity as a base of our framework we evaluate and compare several approaches to improve the segmentation model s architecture and training pipeline including pre-processing techniques encoder backbone deep network types of varying complexity and specialized loss functions to mitigate class imbala...