prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
more broadly our work offers new insights into how people rapidly evaluate act and make suggestions when encountering novel problems and could inform the design of more
flexible
predictive processing
more broadly our work offers new insights into how people rapidly evaluate act and make suggestions when encountering novel problems and could inform the design of more flexible and human-like ai systems that can determine not just how to solve new tasks but whether a task is worth thinking about at all
two companion studies further demonstrate impact at scale gradient-based iterative methods for strategic bidding in energy markets and sobolev-style
training
gradient descent
two companion studies further demonstrate impact at scale gradient-based iterative methods for strategic bidding in energy markets and sobolev-style training of end-to-end optimization proxies using solver-accurate sensitivities
we find that shortcut learning is not localized in specific layers but
distributed
federated learning
we find that shortcut learning is not localized in specific layers but distributed throughout the network
as a result the gromov-wasserstein distance can be used to rank
edges
network structures
as a result the gromov-wasserstein distance can be used to rank edges depending on their criticality with respect to their individual impact on the overall infrastructure and level allowing for prioritizing maintenance emergency planning and enhancing the resilience of the urban transport network
our results underline the potential of generative
models
quantum mechanics
our results underline the potential of generative models as a general-purpose methodology for automated quantum circuit design offering a promising path towards more efficient quantum algorithms and accelerating scientific discovery in the quantum domain
we establish crucial correlations between the mosquito reproduction number mathcal r _ m and the disease
reproduction
reproduction number
we establish crucial correlations between the mosquito reproduction number mathcal r _ m and the disease reproduction number mathcal r _ 0 with the disease dynamics in a multi-patch environment encompassing not only a numerical analysis but also from a theoretical perspective
inference on local variable importance measures for heterogeneous
treatment
treatment effect
inference on local variable importance measures for heterogeneous treatment effects
the electron phonon coupling of a defect characterized by its huang rhys hr factor is a crucial metric determining its excited-state dynamics relevant to defect applications as qubits and
quantum
quantum technologies
the electron phonon coupling of a defect characterized by its huang rhys hr factor is a crucial metric determining its excited-state dynamics relevant to defect applications as qubits and quantum emitters
the era of agentic organization learning to organize with
language
large language
the era of agentic organization learning to organize with language models
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained
bilevel
bilevel optimization
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained bilevel optimization sflcb
to capture the long-term dependency and complex dynamics of
eeg
brain-computer interface
to capture the long-term dependency and complex dynamics of eeg we propose a hybrid encoder combining a mamba-like linear attention channel encoder and a spatiotemporal dynamics model
we apply a fine-tuning-plus-verifier framework in which llm
agents
llm agents
we apply a fine-tuning-plus-verifier framework in which llm agents are equipped with various communication strategies and verification signals from the environment
however its application is often hindered by low textbf
reward
reward density
however its application is often hindered by low textbf reward density in deep search scenarios where agents expend significant exploratory costs for infrequent and often null final rewards
do railway commuters exhibit consistent route choice rationality across
different
human mobility
do railway commuters exhibit consistent route choice rationality across different contexts and time
although it is often regarded as a recent innovation of the
agent
llm agents
although it is often regarded as a recent innovation of the agent era we argue that related practices can be traced back more than twenty years
to address this challenge researchers have recently introduced the concept of
context
context engineering
to address this challenge researchers have recently introduced the concept of context engineering
we give matching lower bounds up to polylogarithmic
factors
mathrm polylog
we give matching lower bounds up to polylogarithmic factors for both results
as an alternative to established hole-based photonic crystal cavities we introduce corrugated triangular dinosaur
photonic
photonic crystal
as an alternative to established hole-based photonic crystal cavities we introduce corrugated triangular dinosaur photonic crystal cavities and develop a tapered quasi loss-free cavity-waveguide interface to adiabatically interconvert bloch and waveguide modes
coupled flow matching builds on two components i an extended gromov-wasserstein optimal
transport
optimal transport
coupled flow matching builds on two components i an extended gromov-wasserstein optimal transport objective that establishes a probabilistic correspondence between data and embeddings and ii a dual-conditional flow-matching network that extrapolates the correspondence to the underlying space
we use real data as well as numerical simulations to
illustrate
empirical application
we use real data as well as numerical simulations to illustrate the performance of the proposed methods
hence we divide the main problem into two sub-problems the antenna activation
sub-problem
wireless networks
hence we divide the main problem into two sub-problems the antenna activation sub-problem and the power allocation sub-problem
we use the database constructed in wasleske baldassare 2024 which contains dwarf galaxies that were selected as
active
massive galaxies
we use the database constructed in wasleske baldassare 2024 which contains dwarf galaxies that were selected as active galaxies by optical spectroscopy infrared colors x-ray brightness and photometric variability
to address these gaps we introduce the waymo open dataset for end-to-end
driving
autonomous driving
to address these gaps we introduce the waymo open dataset for end-to-end driving wod-e2e
this improvement allows the inverse optimal safe control to inherit the standard gain margin of 1 2 inf without requiring prior knowledge of whether f x or u0
acts
gain margin
this improvement allows the inverse optimal safe control to inherit the standard gain margin of 1 2 inf without requiring prior knowledge of whether f x or u0 acts safely on the safety boundary while simultaneously ensuring global asymptotic stability of the resulting safe set
this enables dynamic and real-time tasks that were previously believed to be unattainable by large
vla
vision-language-action vla
this enables dynamic and real-time tasks that were previously believed to be unattainable by large vla models
more broadly they point to an intrinsic neural signature of adaptive brain function marked by efficient yet flexible network organization that may support creative and
adaptive
neural codes
more broadly they point to an intrinsic neural signature of adaptive brain function marked by efficient yet flexible network organization that may support creative and adaptive cognition
on a real robot with 22 degrees of freedom dofs get-use outperforms state-of-the-art methods by 30-60 success rates across three vision-based bimanual mobile
manipulation
robotic manipulation
on a real robot with 22 degrees of freedom dofs get-use outperforms state-of-the-art methods by 30-60 success rates across three vision-based bimanual mobile manipulation tool-usage tasks
6 performance while dtc-vae delivered the most
consistent
consistently outperforms
6 performance while dtc-vae delivered the most consistent his with 92
quasi linear preprocessing time followed by
constant
-time algorithm
quasi linear preprocessing time followed by constant or logarithmic time per output
while meta-reinforcement learning rl agents can attain near bayes-optimal policies they often fail to learn the
compact
continual learning
while meta-reinforcement learning rl agents can attain near bayes-optimal policies they often fail to learn the compact interpretable bayes-optimal belief states
we propose a framework according to which ai
agents
trustworthy ai
we propose a framework according to which ai agents engagement is conditional on appropriate user behaviour
5 varepsilon n for 2 le k le o 1 and any constant
varepsilon
epsilon -approximate
5 varepsilon n for 2 le k le o 1 and any constant varepsilon 0
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
stellar population
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
a core challenge for 1pns is reliably measuring semantic differences between multi-step chains of
thought
thinking traces
a core challenge for 1pns is reliably measuring semantic differences between multi-step chains of thought so we introduce reasoning path divergence rpd a step-level metric that aligns and scores long chain-of-thought solutions to capture differences in intermediate reasoning
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
galaxy mergers trigger starburst activity and
galactic
galactic nuclei
galaxy mergers trigger starburst activity and galactic outflows that enrich the circumgalactic medium profoundly impacting galaxy evolution
machine translation mt is widely employed to address resource scarcity in low-resource
languages
large language
machine translation mt is widely employed to address resource scarcity in low-resource languages by generating synthetic data from high-resource counterparts
emergence of hybrid computational dynamics through
reinforcement
artificial intelligence
emergence of hybrid computational dynamics through reinforcement learning
open-source software is available which implements the
proposed
existing methods
open-source software is available which implements the proposed methods
extensive experiments demonstrate that our method effectively imitates various
categories
extensive experiments
extensive experiments demonstrate that our method effectively imitates various categories of effect information and exhibits outstanding generalization to out-of-domain effects
using a simulation model based on the prisoner s dilemma and quantal response equilibrium we analyze how variations in these learning
rates
learning rates
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
while focused on cooperation our assumptions generalize to any decision-making process involving a
choice
control strategies
while focused on cooperation our assumptions generalize to any decision-making process involving a choice between alternative options
in this work we aim to explain this conflict by exploring how language models manipulate
numbers
large language models llms
in this work we aim to explain this conflict by exploring how language models manipulate numbers and quantify the lower bounds of accuracy of these mechanisms
state-space models ssms are a widely used tool in time
series
time series
state-space models ssms are a widely used tool in time series analysis
our analysis focuses on cases where this game qualifies as a markov potential game mpg a class of games where we can provide an alignment guarantee under a structural assumption on the human s value function any decision by the
agent
reinforcement learning
our analysis focuses on cases where this game qualifies as a markov potential game mpg a class of games where we can provide an alignment guarantee under a structural assumption on the human s value function any decision by the agent to act more autonomously that benefits itself cannot harm the human s value
furthermore we introduce vnia visual narrative intent alignment a
multimodal
vision-language models vlms
furthermore we introduce vnia visual narrative intent alignment a multimodal dataset specifically created to facilitate the development and evaluation of vlm steering techniques
our findings highlight the importance of designing interactions with
generative
generative ai
our findings highlight the importance of designing interactions with generative ai systems as reflective thought partners that complement human strengths and augment creative processes
our results imply that quantum fluctuations have a measurable influence on selecting the
ground
ground state
our results imply that quantum fluctuations have a measurable influence on selecting the ground state of a system out of competing ordered magnetic phases at low temperature
consider the setting in which a researcher is interested in the causal effect of a treatment z on a duration
time
treatment effect
consider the setting in which a researcher is interested in the causal effect of a treatment z on a duration time t which is subject to right censoring
we also present a range of illustrative examples including correlated versions of erd h o s-r enyi and inhomogeneous random graph models and
dynamic
correlation network
we also present a range of illustrative examples including correlated versions of erd h o s-r enyi and inhomogeneous random graph models and dynamic networks
for task assignment we present two multi-objective algorithms non-dominated sorting genetic
algorithm
objective function
for task assignment we present two multi-objective algorithms non-dominated sorting genetic algorithm nsga and adaptive large neighborhood search alns
this formulates the problem as a co-optimization setup to i optimize the data processing and ii optimally
allocate
optimization problem
this formulates the problem as a co-optimization setup to i optimize the data processing and ii optimally allocate the computing resources
yet an important question still remains are video models ready to serve as zero-shot reasoners in challenging visual
reasoning
multimodal reasoning
yet an important question still remains are video models ready to serve as zero-shot reasoners in challenging visual reasoning scenarios
large language models llms are widely used in
generative
natural language
large language models llms are widely used in generative applications such as chatting code generation and reasoning
instrumental variable methods are fundamental to causal inference when
treatment
treatment regimes
instrumental variable methods are fundamental to causal inference when treatment assignment is confounded by unobserved variables
meanwhile it provides a unified theoretical explanation for classic topological characteristics such as small-world networks and
scale-free
scale-free networks
meanwhile it provides a unified theoretical explanation for classic topological characteristics such as small-world networks and scale-free networks
its success is partly attributed to conditioning policies on large fixed
context
context length
its success is partly attributed to conditioning policies on large fixed context length
we present a generalization of this problem where inspired by biological considerations the
food
ecological communities
we present a generalization of this problem where inspired by biological considerations the food web has weighted edges to represent the importance of predator-prey relationships
we present a general strategy for turning
generative
generative models
we present a general strategy for turning generative models into candidate solution samplers for batch bayesian optimization bo
the report is organized around three core capabilities required for brain emulation recording brain function neural dynamics mapping brain structure connectomics and
emulation
surrogate brain
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
prior work has typically compared systems using a single representational similarity
metric
abstract representations
prior work has typically compared systems using a single representational similarity metric yet each captures only one facet of representational structure
this theory further promotes nonlinear transport as a probe of
geometric
third-order nonlinear transport
this theory further promotes nonlinear transport as a probe of geometric effects and phase transitions in quantum materials
the discriminator is trained using a combination of real-world data and simulation data executed by the agent which is designed to train a
policy
imitation learning
the discriminator is trained using a combination of real-world data and simulation data executed by the agent which is designed to train a policy that generates realistic motion trajectories that match the statistical properties of human motion
artificial intelligence systems based on large language
models
ai use
artificial intelligence systems based on large language models llms can now generate coherent text music and images yet they operate without a persistent state each inference reconstructs context from scratch
simulations showed that grassmannian signaling provides competitive bit error rates ber at low signal-to-noise
ratio
signal-to-noise ratio
simulations showed that grassmannian signaling provides competitive bit error rates ber at low signal-to-noise ratio snr regimes with low probability of detection at the unintended receiver compared to coherent schemes that use qpsk or qam modulation formats and need pilots to perform channel estimation
the basic reproduction number mathcal r_0 calculated from models that assume homogeneous mixing or single-layer contact structures have limited
applicability
basic reproduction
the basic reproduction number mathcal r_0 calculated from models that assume homogeneous mixing or single-layer contact structures have limited applicability to complex social systems
agent-based simulations reproduce these patterns under controlled
dynamics
traffic dynamics
agent-based simulations reproduce these patterns under controlled dynamics and a simple analytical decomposition clarifies why network interactions explain a large share of cross-sectional variance when only macro-level counts are available but much less once recent town-level case histories are included
we find that simply imposing the m_ rm bh - sigma_ star condition is sufficient to generate a
fraction
stellar population
we find that simply imposing the m_ rm bh - sigma_ star condition is sufficient to generate a fraction of quenched galaxies consistent with current data including the newest ones from euclid
unsupervised learning is therefore a natural approach for exploring the design of
biological
neural codes
unsupervised learning is therefore a natural approach for exploring the design of biological neural networks and their computations
the ac optimal power flow ac-opf problem is central to
power
optimal power
the ac optimal power flow ac-opf problem is central to power system operation but challenging to solve efficiently due to its nonconvex and nonlinear nature
across three real-world tasks and two embodiments hi-ors fine-tunes a pi-base policy to master contact-rich
manipulation
manipulation ordering
across three real-world tasks and two embodiments hi-ors fine-tunes a pi-base policy to master contact-rich manipulation in just 1
we provide the first multi-wavelength polarisation decomposed characterisation of the principal modes of a
photonic
integrated photonics
we provide the first multi-wavelength polarisation decomposed characterisation of the principal modes of a photonic lantern
modeling object attention in mobile ar for intrinsic
cognitive
object recall
modeling object attention in mobile ar for intrinsic cognitive security
scalable solid state single-photon sources spss with triggered single-photon emission rates exceeding a few ghz would aid in the wide technological adoption of photonic
quantum
quantum emitters
scalable solid state single-photon sources spss with triggered single-photon emission rates exceeding a few ghz would aid in the wide technological adoption of photonic quantum technologies
finally we outline open problems in the field and present a roadmap for integrating tf-qkd into scalable quantum
networks
quantum advantage
finally we outline open problems in the field and present a roadmap for integrating tf-qkd into scalable quantum networks underscoring its central role in the future quantum internet
morphology-aware graph reinforcement learning for tensegrity
robot
motion planning
morphology-aware graph reinforcement learning for tensegrity robot locomotion
formulas for calculating the mean square errors and the spectral characteristics of the optimal linear estimate of the functional are derived in the case where the
spectral
spectral density
formulas for calculating the mean square errors and the spectral characteristics of the optimal linear estimate of the functional are derived in the case where the spectral density matrices are exactly known
i develop a nonparametric framework for identifying
spatial
spatial treatment
i develop a nonparametric framework for identifying spatial boundaries of treatment effects without imposing parametric functional form restrictions
identifying geometric third-order nonlinear
transport
third-order nonlinear transport
identifying geometric third-order nonlinear transport in disordered materials
it is not possible for example to get o 1
query
query time
it is not possible for example to get o 1 query time using nv o n bits of space when v theta log n and assuming the word ram model with o log n -bit words
these results confirm the potential of lightweight ai-based
csi
channel state information csi
these results confirm the potential of lightweight ai-based csi prediction to effectively mitigate channel aging and enhance link adaptation in tdd 5g systems
while diffusion language models dlms enable fine-grained refinement their practical controllability
remains
language agents
while diffusion language models dlms enable fine-grained refinement their practical controllability remains fragile
in this work we present a clear and robust morphological analysis of a sample of 190 galaxies at z 6 demonstrating that distinct
bulge
massive stars
in this work we present a clear and robust morphological analysis of a sample of 190 galaxies at z 6 demonstrating that distinct bulge and disk components were already beginning to emerge during this early epoch
in this paper we conduct a unified sample complexity analysis of
zeroth-order
zeroth-order methods
in this paper we conduct a unified sample complexity analysis of zeroth-order methods across gradient estimators with different search directions
more broadly they point to an intrinsic neural signature of adaptive brain function marked by efficient yet flexible network organization that may support creative and
adaptive
artificial neural
more broadly they point to an intrinsic neural signature of adaptive brain function marked by efficient yet flexible network organization that may support creative and adaptive cognition
4 the siren x-ray and optical emissions take over each other twice per
cycle
vi emission
4 the siren x-ray and optical emissions take over each other twice per cycle possibly with two different peak x-ray fluxes within one cycle
in this study we first prove that the density-ratio estimation method proposed in
lin
ate estimation
in this study we first prove that the density-ratio estimation method proposed in lin et al
we quantify the dependence of magnetic fields on star formation activity including both
regular
magnetic field
we quantify the dependence of magnetic fields on star formation activity including both regular and starburst galaxies
this paper proposes a new method for estimating conditional
average
treatment effect boundaries
this paper proposes a new method for estimating conditional average treatment effects cate in randomized experiments
our formation planner is a two-step optimization problem that identifies desired relative
robot
multi-robot collaboration
our formation planner is a two-step optimization problem that identifies desired relative robot positions
the structure of relation decoding linear operators in large
language
language models
the structure of relation decoding linear operators in large language models
neuronmm high-performance matrix multiplication for llm
inference
models llms
neuronmm high-performance matrix multiplication for llm inference on aws trainium
our results constitute a substantial development on existing corrections to mean-field theory for infectious
individuals
infectious individuals
our results constitute a substantial development on existing corrections to mean-field theory for infectious individuals in sis processes and provide an in-depth characterization of how structural randomness in networks affects the dynamical trajectories of infectious diseases on networks
our approach culminates in the successful
deployment
llm post-training
our approach culminates in the successful deployment of llms into a live large-scale online ai service
unlike autoregressive captioning the strength of the
visual
vision-language models vlms
unlike autoregressive captioning the strength of the visual learning signal in mdc does not depend on each token s position in the sequence reducing the need for auxiliary objectives
we further reveal that quantum jump induce quasicycles whose amplitude
scales
super-heisenberg scaling
we further reveal that quantum jump induce quasicycles whose amplitude scales inversely with the square root of the system size
a prevailing theory in cognitive neuroscience proposes that the human brain operates through hierarchical predictive processing which underlies flexible
cognition
human brain
a prevailing theory in cognitive neuroscience proposes that the human brain operates through hierarchical predictive processing which underlies flexible cognition and behavior by integrating multimodal sensory signals
simulation studies demonstrate the finite-sample performance of the
proposed
numerical experiments
simulation studies demonstrate the finite-sample performance of the proposed method and two empirical applications illustrate its practical utility
the external medium also influences the evolution of circumstellar disks and protostellar outflows with the high-density external medium disks grow rapidly but their mass becomes smaller relative to the
protostellar
circumgalactic medium
the external medium also influences the evolution of circumstellar disks and protostellar outflows with the high-density external medium disks grow rapidly but their mass becomes smaller relative to the protostellar mass and the outflow is sustained over a long duration
the new technique is executed with a spatially filtered collimated gaussian beam that is propagated through computer controlled rotating prpp to introduce the lab based turbulence on the
beam
beam shaping
the new technique is executed with a spatially filtered collimated gaussian beam that is propagated through computer controlled rotating prpp to introduce the lab based turbulence on the beam profile