prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
this way prior domain knowledge can be incorporated into the learning process and the learned
driving
autonomous driving
this way prior domain knowledge can be incorporated into the learning process and the learned driving behaviour can be constrained more easily
we show that if the system is controllable then incorporating this as prior knowledge does not relax the
conditions
control systems
we show that if the system is controllable then incorporating this as prior knowledge does not relax the conditions required for data-driven stabilization
leveraging this equivalence we propose a novel regularization method for
policy
policy learning
leveraging this equivalence we propose a novel regularization method for policy learning
we investigate the chemical content of 22 well-studied massive protostars from the sofia massive soma star formation survey aiming to identify correlations between
chemical
star formation rates
we investigate the chemical content of 22 well-studied massive protostars from the sofia massive soma star formation survey aiming to identify correlations between chemical and physical parameters
because it operates purely at inference time it applies uniformly across various dlms and naturally extends to
diverse
llm inference
because it operates purely at inference time it applies uniformly across various dlms and naturally extends to diverse supervision sources
this paper presents a comprehensive cross-platform evaluation of reasoning capabilities in contemporary
foundation
foundation models
this paper presents a comprehensive cross-platform evaluation of reasoning capabilities in contemporary foundation models establishing an infrastructure-agnostic benchmark across three computational paradigms hpc supercomputing marenostrum 5 cloud platforms nebius ai studio and university clusters a node with eight h200 gpus
in real-world environments ai systems often face unfamiliar scenarios without labeled data creating a major challenge for conventional scene
understanding
vision-language models vlms
in real-world environments ai systems often face unfamiliar scenarios without labeled data creating a major challenge for conventional scene understanding models
we establish the convergence of the sequence generated by tkma to a fixed
point
convergence guarantees
we establish the convergence of the sequence generated by tkma to a fixed point of t
the report is organized around three core capabilities required for brain emulation recording brain function neural dynamics mapping brain structure connectomics and
emulation
electroencephalography eeg
the report is organized around three core capabilities required for brain emulation recording brain function neural dynamics mapping brain structure connectomics and emulation and embodiment computational neuroscience
the turbulence impacted beam is then propagated through vortex phase plate vpp to shape that
beam
beam shaping
the turbulence impacted beam is then propagated through vortex phase plate vpp to shape that beam into different topological charged laguerre gaussian modes and further collimated using two lens based configuration
we motivate and develop four idealized galaxy cgm
spatial-kinematic
galaxy cgm
we motivate and develop four idealized galaxy cgm spatial-kinematic structures based on empirical data and theoretical predictions 1 a rotating galactic disk extra-planar gas 2 a static or dynamic spherical halo 3 an outflowing bi-polar galactic wind and 4 an inward spiraling flared planar accretion
we present a generalization of the so-called fractional stirap f-stirap demonstrating precise control over the mixing ratio of
quantum
quantum emitters
we present a generalization of the so-called fractional stirap f-stirap demonstrating precise control over the mixing ratio of quantum states in the wave packet
the bounds are the solutions to optimization problems for which we derive a
computationally
optimization problem
the bounds are the solutions to optimization problems for which we derive a computationally tractable dual formulation
we compile multiwavelength data from the radio to the gamma-ray regimes for all sources including
multiwavelength
gamma -ray
we compile multiwavelength data from the radio to the gamma-ray regimes for all sources including multiwavelength spectral indices and redshifts
large language models llms face significant inference latency challenges stemming from their autoregressive design and
large
large language
large language models llms face significant inference latency challenges stemming from their autoregressive design and large size
to address this tension a coach-like llm environment was developed that embodies divergent and convergent
thinking
llm reasoning
to address this tension a coach-like llm environment was developed that embodies divergent and convergent thinking personas as two complementary processes
through this survey we aim to provide a foundational understanding of how ai techniques are reshaping thermal science and guiding future research in
nanoscale
heat conduction
through this survey we aim to provide a foundational understanding of how ai techniques are reshaping thermal science and guiding future research in nanoscale heat transfer
our solutions assume a constant star formation efficiency a constant mass-loading factor and that the yields are linearly dependent on the interstellar medium
abundance
star formation rates
our solutions assume a constant star formation efficiency a constant mass-loading factor and that the yields are linearly dependent on the interstellar medium abundance with the option of a saturation of the yields at high metallicity
although various high power-efficiency waveforms have been proposed most are incompatible with the cp-ofdma framework and remain ineffective in multi-user
downlink
uplink communication
although various high power-efficiency waveforms have been proposed most are incompatible with the cp-ofdma framework and remain ineffective in multi-user downlink transmissions
learning to plan schedule with reinforcement-learned bimanual
robot
imitation learning
learning to plan schedule with reinforcement-learned bimanual robot skills
kernel density estimation kde is a cornerstone of
nonparametric
density estimation
kernel density estimation kde is a cornerstone of nonparametric statistics yet it remains sensitive to bandwidth choice boundary bias and computational inefficiency
we provide a sharp analytical characterization of the stochastic steady states and of the
transition
game theory
we provide a sharp analytical characterization of the stochastic steady states and of the transition dynamics across nash equilibria and employ simulations to illustrate these results in special cases
we evaluate our approach on real-world datasets and demonstrate that integrating interventional constraints not only improves model accuracy and ensures consistency with established findings making models more explainable but also facilitates the discovery of new
causal
interventional constraints
we evaluate our approach on real-world datasets and demonstrate that integrating interventional constraints not only improves model accuracy and ensures consistency with established findings making models more explainable but also facilitates the discovery of new causal relationships that would otherwise be costly to identify
reliable estimation of network-wide traffic states is essential for urban
traffic
reliable estimation
reliable estimation of network-wide traffic states is essential for urban traffic management
these features suggest that the gas forming the cold front is flowing in the plane oriented along the los supporting an offset merger scenario in which the main
cluster
star formation
these features suggest that the gas forming the cold front is flowing in the plane oriented along the los supporting an offset merger scenario in which the main cluster has passed in front of the subcluster and induced rotation of the core gas in the plane perpendicular to the sky
we then discuss how the geometric primitives naturally embed nullspace structures into the controllers that can be exploited for introducing secondary
control
optimal control
we then discuss how the geometric primitives naturally embed nullspace structures into the controllers that can be exploited for introducing secondary control objectives
simulating the dynamics of quantum impurity
models
quantum computing
simulating the dynamics of quantum impurity models remains a fundamental challenge due to the complex memory effects that arise from system-environment interactions
we track the stellar populations in bulges back in time to their birth location classifying them as bulge-born in-situ and disk-born
stars
massive stars
we track the stellar populations in bulges back in time to their birth location classifying them as bulge-born in-situ and disk-born stars and accreted
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
we evaluate 15 foundation models across 79 problems spanning eight academic
domains
world models
we evaluate 15 foundation models across 79 problems spanning eight academic domains physics mathematics chemistry economics biology statistics calculus and optimization through three experimental phases 1 baseline establishment six models mixtral-8x7b phi-3 llama 3
wigner negativity and genuine multipartite entanglement gme are key nonclassical
resources
entanglement entropy
wigner negativity and genuine multipartite entanglement gme are key nonclassical resources that enable computational advantages and broader quantum-information tasks
optimization methods that make use of derivatives of the objective function up to order p 2 are called
tensor
tensor product
optimization methods that make use of derivatives of the objective function up to order p 2 are called tensor methods
giving humans an ai fact-verification assistant further improves their accuracy but the type of
assistance
ai assistance
giving humans an ai fact-verification assistant further improves their accuracy but the type of assistance matters
this theory further promotes nonlinear transport as a probe of geometric effects and phase transitions in
quantum
quantum technologies
this theory further promotes nonlinear transport as a probe of geometric effects and phase transitions in quantum materials
viruses are microscopic infectious agents that require a
host
infectious individuals
viruses are microscopic infectious agents that require a host cell for replication
variational quantum algorithms vqas face an inherent trade-off between expressivity and trainability deeper circuits can represent richer
states
quantum advantage
variational quantum algorithms vqas face an inherent trade-off between expressivity and trainability deeper circuits can represent richer states but suffer from noise accumulation and barren plateaus while shallow circuits remain trainable and implementable but lack expressive power
evaluating policy effects under network interference without network information a transfer
learning
policy learning
evaluating policy effects under network interference without network information a transfer learning approach
this approach also learns the joint distribution between spatial and non-spatial features exploiting
spatial
local gaussian correlation
this approach also learns the joint distribution between spatial and non-spatial features exploiting spatial autocorrelations
our comprehensive evaluation of open-weight and proprietary reasoning and non-reasoning
vlms
models vlms
our comprehensive evaluation of open-weight and proprietary reasoning and non-reasoning vlms reveals that most models perform near chance and even the best lag far behind human accuracy on physically irreversible processes e
variational mean field approximations tend to struggle with contemporary overparametrized
deep
deep learning
variational mean field approximations tend to struggle with contemporary overparametrized deep neural networks
it groups low-level perceptual features into high-level object representations stores those objects efficiently and compositionally in memory and supports human
reasoning
human cognition
it groups low-level perceptual features into high-level object representations stores those objects efficiently and compositionally in memory and supports human reasoning about individual object instances
identification and debiased learning of causal effects with
general
debiased machine learning
identification and debiased learning of causal effects with general instrumental variables
high-speed generation processing and detection of polarization bell-states is therefore critical for
quantum
quantum dot
high-speed generation processing and detection of polarization bell-states is therefore critical for quantum technology
furthermore this approach is generalizable beyond preference-based feedback to general types of reward signals and
loss
loss functions
furthermore this approach is generalizable beyond preference-based feedback to general types of reward signals and loss functions
the resulting synthetic corpus is used for continual pretraining with an interleaved
curriculum
reasoning curriculum
the resulting synthetic corpus is used for continual pretraining with an interleaved curriculum schedule aligning learning across both content and cognitive dimensions
the proposed approach integrates pre-trained vision
transformers
vision transformers
the proposed approach integrates pre-trained vision transformers and large language models to align visual semantics with natural language descriptions enhancing contextual comprehension
finally we demonstrate how this could be used in the field by having a
robot
mobile robots
finally we demonstrate how this could be used in the field by having a robot attempt a gap-crossing task with and without inflating its actuators
overall our model provides a framework for investigating the importance of cognitive and
social
human disturbance
overall our model provides a framework for investigating the importance of cognitive and social structures in determining human-environment dynamics
joint uplink and downlink resource allocation and antenna activation for pinching
antenna
uplink communication
joint uplink and downlink resource allocation and antenna activation for pinching antenna systems
our work provides a scalable method to overcome a measurement bottleneck in
cognitive
cognitive science
our work provides a scalable method to overcome a measurement bottleneck in cognitive science and demonstrates that foundation models can learn a representational geometry that is functionally relevant for modeling key aspects of human cognition such as categorization
many believe that intracortical axons conduct
signals
brain decoding
many believe that intracortical axons conduct signals too slowly to bring the contextual information from receptive fields of other neurons
communication impact is quantied by a capacity-motivated lower bound obtained from the linear minimum mean-squared error
error
signal-to-noise ratio
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 calibration method the plate design the mathematical description of the error system as well as the identification of the parameters are
described
calibration plate
the calibration method the plate design the mathematical description of the error system as well as the identification of the parameters are described in detail
this study not only enriches the theoretical framework of evolutionary game theory but also provides a foundation for the management of ecological
systems
collective systems
this study not only enriches the theoretical framework of evolutionary game theory but also provides a foundation for the management of ecological systems and the design of cooperative mechanisms in society
exogenous processes -- that affect the individual and are not derived from social
interactions
collective systems
exogenous processes -- that affect the individual and are not derived from social interactions -- even if unbiased have a role in supporting cooperation over defection and this role has been largely overlooked in the context of network-based interactions
we then discuss how the geometric primitives naturally embed nullspace structures into the
controllers
optimal control
we then discuss how the geometric primitives naturally embed nullspace structures into the controllers that can be exploited for introducing secondary control objectives
our initial approach uses solve_bvp to approximate optimal
control
linear control
our initial approach uses solve_bvp to approximate optimal control trajectories
these findings highlight the potential of shared control strategies to balance
stability
control strategy
these findings highlight the potential of shared control strategies to balance stability efficiency and driver acceptance
here we introduce the geohabnet r package which evaluates the potential importance of locations for the spread of
species
ecological interactions
here we introduce the geohabnet r package which evaluates the potential importance of locations for the spread of species through habitat landscapes
for instances with more than 100 jobs exact methods such as mip and dynamic
programming
efficiently solving
for instances with more than 100 jobs exact methods such as mip and dynamic programming become computationally intractable
however theoretical studies often rely on simplified
architectures
theoretical findings
however theoretical studies often rely on simplified architectures e
gradient flow sampler-based distributionally
robust
gradient flow
gradient flow sampler-based distributionally robust optimization
with enough data optimal performance is attained through model s specialisation towards the target but it can be hard to reach for training algorithms which get attracted by sub-optimal
solutions
reinforcement learning
with enough data optimal performance is attained through model s specialisation towards the target but it can be hard to reach for training algorithms which get attracted by sub-optimal solutions predicted by the theory
8 probability that arbitrarily advanced technology could induce vacuum decay if our
vacuum
vacuum metastable
8 probability that arbitrarily advanced technology could induce vacuum decay if our vacuum is metastable
large language models llms are widely used in generative
applications
llm agents
large language models llms are widely used in generative applications such as chatting code generation and reasoning
we study a coordinated multi-point comp transmission where two
base
base station
we study a coordinated multi-point comp transmission where two base stations bss each supported by a pinching antenna system pass are deployed to jointly serve communication users under spatial division multiple access sdma technology
we benchmark the method using mock datasets representative of high-galactic-latitude regions incorporating realistic
gaia
galactic disk
we benchmark the method using mock datasets representative of high-galactic-latitude regions incorporating realistic gaia parallax uncertainties and polarization expected from the upcoming pasiphae survey
in this paper we systematically evaluate llms reasoning
capabilities
reasoning capabilities
in this paper we systematically evaluate llms reasoning capabilities in the normative domain from both logical and modal perspectives
in third-order nonlinear transport a voltage can be measured in response to the cube of a driving current as a result of the quantum
geometric
third-order nonlinear
in third-order nonlinear transport a voltage can be measured in response to the cube of a driving current as a result of the quantum geometric effects which has attracted tremendous attention
the retrieved complex refractive index reveals distinct
modes
waveguide modes
the retrieved complex refractive index reveals distinct modes at 0
here we propose an expression of mathcal r_0 in the context of
multiplex
multiplex networks
here we propose an expression of mathcal r_0 in the context of multiplex networks enabling the analysis of disease transmission across multiple social layers
there are two major approaches in policy learning the empirical
welfare
policy learning
there are two major approaches in policy learning the empirical welfare maximization ewm approach and the plug-in approach
classical approaches to sample selection rely on strong parametric distributional
assumptions
nonparametric identification
classical approaches to sample selection rely on strong parametric distributional assumptions which may be restrictive in practice
we investigate the thermal evolution of 3i atlas the third macroscopic
interstellar
dwarf galaxies
we investigate the thermal evolution of 3i atlas the third macroscopic interstellar object discovered on 2025 july 1
in this study we first prove that the density-ratio
estimation
density ratio
in this study we first prove that the density-ratio estimation method proposed in lin et al
participants in our study n 13 found the system highly immersive and engaging with narrator and audio most impactful while also highlighting areas for improvement in latency and
image
user experience
participants in our study n 13 found the system highly immersive and engaging with narrator and audio most impactful while also highlighting areas for improvement in latency and image resolution
efficiency without cognitive change evidence from human interaction with
narrow
task performance
efficiency without cognitive change evidence from human interaction with narrow ai systems
we demonstrate the practical value of our approach through applications across economics biology and machine
learning
machine learning
we demonstrate the practical value of our approach through applications across economics biology and machine learning benchmarks
to enhance efficiency we further develop vimogen-light a distilled variant that eliminates video
generation
image generation
to enhance efficiency we further develop vimogen-light a distilled variant that eliminates video generation dependencies while preserving strong generalization
inspired by predictive coding in neuroscience--which suggests that the brain predicts sensory inputs as a neural implementation of bayesian inference--and by auxiliary predictive objectives in deep rl we investigate whether integrating self-supervised
predictive
brain decoding
inspired by predictive coding in neuroscience--which suggests that the brain predicts sensory inputs as a neural implementation of bayesian inference--and by auxiliary predictive objectives in deep rl we investigate whether integrating self-supervised predictive coding modules into meta-rl can facilitate learning of bayes-optimal representations
we evaluate the model in simulation with up to 14
qubits
qubit readout
we evaluate the model in simulation with up to 14 qubits on sentiment analysis mnist permuted mnist copying memory and language modeling adopting projective measurements as a limiting case to obtain mid-circuit readouts while maintaining a coherent recurrent quantum memory
smbhs are evolved following the stellar mass growth of their host galaxies by assigning an accretion
rate
stellar population
smbhs are evolved following the stellar mass growth of their host galaxies by assigning an accretion rate at each redshift from the empirical eddington ratio distributions and duty cycles
through comprehensive experimentation across six diverse tasks and utilizing six distinct
llms
models llms
through comprehensive experimentation across six diverse tasks and utilizing six distinct llms our methodology demonstrates remarkable results achieving speeds up to 5
this result means that fast transport particularly in dense clumps is necessary for simulations to agree with the dearth of observations of
gamma
gamma -ray
this result means that fast transport particularly in dense clumps is necessary for simulations to agree with the dearth of observations of gamma -ray emission from diffuse gas like the cgm and icm
notably under a similar performance guarantee as in our
tree
tree edit distance
notably under a similar performance guarantee as in our tree embedding algorithms i
finally we focus on polyhedral norms and show that normal
curves
normal curves
finally we focus on polyhedral norms and show that normal curves have controls that locally take values in a single face of a sphere with respect to the norm
while modern large language models llms are increasingly used to model neural responses to
language
recurrent neural networks
while modern large language models llms are increasingly used to model neural responses to language their internal representations are highly entangled mixing information about lexicon syntax meaning and reasoning
we investigate the chemical content of 22 well-studied massive protostars from the sofia massive soma star formation survey aiming to identify correlations between
chemical
star formation
we investigate the chemical content of 22 well-studied massive protostars from the sofia massive soma star formation survey aiming to identify correlations between chemical and physical parameters
yet robust decisions under uncertainty still rely on capabilities that
current
ai systems
yet robust decisions under uncertainty still rely on capabilities that current ai lacks domain knowledge not captured by data long horizon context and reasoning grounded in the physical world
we evaluate the method in simulated 2d environments with varying road capacities and
traffic
autonomous driving
we evaluate the method in simulated 2d environments with varying road capacities and traffic conditions demonstrating high task completion rates and robust behavior even under congestion
however many existing methods suffer from premature convergence limited exploration and performance degradation in
large-scale
efficiently solving
however many existing methods suffer from premature convergence limited exploration and performance degradation in large-scale search spaces
our work identifies how enhanced coulomb interactions in two dimensions can stabilize this phase at significantly higher temperatures proposing
promising
quantum materials
our work identifies how enhanced coulomb interactions in two dimensions can stabilize this phase at significantly higher temperatures proposing promising material candidates for observing these collective states
this resolution allows us to perform the first star-by-star galaxy
simulation
host galaxy
this resolution allows us to perform the first star-by-star galaxy simulation which resolves individual stars in the milky way galaxy
the simulated galaxies reproduce the observed stellar-to-halo mass mass--metallicity and size--mass relations yielding stellar
masses
stellar mass function
the simulated galaxies reproduce the observed stellar-to-halo mass mass--metallicity and size--mass relations yielding stellar masses of 10 6-10 8 m_ odot and metallicities consistent with those of local group dwarf galaxies
planning is a critical component of end-to-end
autonomous
autonomous driving
planning is a critical component of end-to-end autonomous driving
in this paper we systematically evaluate llms reasoning
capabilities
reasoning tasks
in this paper we systematically evaluate llms reasoning capabilities in the normative domain from both logical and modal perspectives
allowing the order of quantum operations to exist in superposition is known to
open
quantum networks
allowing the order of quantum operations to exist in superposition is known to open new routes for thermodynamic tasks
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
understanding how the human brain progresses from processing simple linguistic inputs to performing high-level
reasoning
human cognition
understanding how the human brain progresses from processing simple linguistic inputs to performing high-level reasoning is a fundamental challenge in neuroscience
inspired by predictive coding in neuroscience--which suggests that the brain predicts sensory inputs as a neural implementation of bayesian inference--and by auxiliary predictive objectives in deep rl we investigate whether integrating self-supervised
predictive
recurrent neural
inspired by predictive coding in neuroscience--which suggests that the brain predicts sensory inputs as a neural implementation of bayesian inference--and by auxiliary predictive objectives in deep rl we investigate whether integrating self-supervised predictive coding modules into meta-rl can facilitate learning of bayes-optimal representations