prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
we further extend these analyses to two broad families of activation functions and deep feedforward architectures demonstrating that
abstract
abstract representations
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 summary this work advances the modeling of epidemics to a more local scope offering a more expressive tool for epidemiological research and
public
infectious individuals
in summary this work advances the modeling of epidemics to a more local scope offering a more expressive tool for epidemiological research and public health planning
rules including stochastic gradient descent sgd feedback alignment fa direct feedback
alignment
dynamical systems
rules including stochastic gradient descent sgd feedback alignment fa direct feedback alignment dfa and kolen-pollack kp emerge naturally as limiting cases of the dynamics
we conduct comprehensive experiments with llama 2-7b llama 2-13b and mistral 7b models on mathematical
reasoning
large language models llms
we conduct comprehensive experiments with llama 2-7b llama 2-13b and mistral 7b models on mathematical reasoning coding and summarization tasks
beyond imitation constraint-aware trajectory generation with flow matching for
end-to-end
autonomous driving
beyond imitation constraint-aware trajectory generation with flow matching for end-to-end autonomous driving
here we explore topological effects on photon-pair generation via spontaneous parametric down-conversion spdc in nonlinear
waveguide
waveguide modes
here we explore topological effects on photon-pair generation via spontaneous parametric down-conversion spdc in nonlinear waveguide arrays both theoretically and experimentally
we find that for different types of network architectures and for both visual or neuronal stimuli these cross-supervising networks learn semantic representations that are easily decodable and that decoding accuracy is comparable to
supervised
deep learning
we find that for different types of network architectures and for both visual or neuronal stimuli these cross-supervising networks learn semantic representations that are easily decodable and that decoding accuracy is comparable to supervised networks -- both at the level of single networks and the ensemble
infoflow reinforcing search agent via reward
density
reinforcement learning
infoflow reinforcing search agent via reward density optimization
a common approach to addressing the challenge involves generating synthetic data for the minority group and then training classification models with both observed and
synthetic
synthetic data
a common approach to addressing the challenge involves generating synthetic data for the minority group and then training classification models with both observed and synthetic data
quantum simulation of actinide chemistry towards scalable algorithms on
trapped
trapped ions
quantum simulation of actinide chemistry towards scalable algorithms on trapped ion quantum computers
we present a generalization of the so-called fractional stirap f-stirap demonstrating precise control over the mixing ratio of
quantum
quantum technologies
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 proposed approach is evaluated across kronecker and weichselberger
channel
channel estimation
the proposed approach is evaluated across kronecker and weichselberger channel models with three distinct pilot probing schemes
by distributing and recollecting the quantum state with an entanglement-distribution operation the scan rate scales as
n
quantum algorithm
by distributing and recollecting the quantum state with an entanglement-distribution operation the scan rate scales as n 2 m 1 while thermal excitation is the dominant background significantly outperforming classical single-cavity methods under matched conditions
moreover clustering cortical regions using snf-derived similarity scores reveals a clearer hierarchical organization that aligns closely with established anatomical and
functional
functional connectivity
moreover clustering cortical regions using snf-derived similarity scores reveals a clearer hierarchical organization that aligns closely with established anatomical and functional hierarchies of the visual cortex-surpassing the correspondence achieved by individual metrics
our formation planner is a two-step optimization
problem
optimal control
our formation planner is a two-step optimization problem that identifies desired relative robot positions
this work extends the balanced network paradigm offering insights into how cortical circuits could maintain robust dynamics via
synaptic
functional connectivity
this work extends the balanced network paradigm offering insights into how cortical circuits could maintain robust dynamics via synaptic adaptation
fine-tuned language models for domain-specific
summarization
natural language processing
fine-tuned language models for domain-specific summarization and tagging
sensitivity analysis for treatment effects in difference-in-differences models using
riesz
riesz representer
sensitivity analysis for treatment effects in difference-in-differences models using riesz representation
the tidal torque theory ttt predicts that galaxy spins are correlated with the surrounding tidal field reflecting how
angular
angular momentum
the tidal torque theory ttt predicts that galaxy spins are correlated with the surrounding tidal field reflecting how angular momentum is acquired during structure formation
they have been observed in many star clusters most of them old globular
clusters
bulge stars
they have been observed in many star clusters most of them old globular clusters populating the milky way and other satellite galaxies
these results suggest that llms can serve as practical proxies for otherwise unrevealed human thinking traces enabling label-only corpora to be extended into thinking-trace-augmented resources that enhance the reliability of
llm
proxy llms
these results suggest that llms can serve as practical proxies for otherwise unrevealed human thinking traces enabling label-only corpora to be extended into thinking-trace-augmented resources that enhance the reliability of llm raters
finally we analyze the generalization performance of a gradient-based
meta-reinforcement
reward models
finally we analyze the generalization performance of a gradient-based meta-reinforcement learning algorithm
in this paper we combine adaptation and control barrier functions into a real-time control architecture that guarantees stability ensures
control
predictive control
in this paper we combine adaptation and control barrier functions into a real-time control architecture that guarantees stability ensures control performance and remains safe even with the parametric uncertainties
artificial intelligence in medicine is built to
serve
artificial intelligence
artificial intelligence in medicine is built to serve the average patient
the koopman operator theory is introduced to transform the nonlinear dynamical differential equations for the dynamic evolution of complex networks into linear
systems
complex systems
the koopman operator theory is introduced to transform the nonlinear dynamical differential equations for the dynamic evolution of complex networks into linear systems for solving
at high pressures nuclear quantum effects involving both hydrogen molecules and the water lattice become dominant giving rise to a dual-lattice
quantum
quantum materials
at high pressures nuclear quantum effects involving both hydrogen molecules and the water lattice become dominant giving rise to a dual-lattice quantum system
maintaining physical activity is essential for older
adults
older adults
maintaining physical activity is essential for older adults health and well-being yet participation remains low
and confirms that controlling the extent of
non-linear
non-linear link
and confirms that controlling the extent of non-linear in preferential attachment is key to achieving a better fit to the real network s degree distribution pattern
it introduces two key capabilities automated feedback generation using a fine-tuned large
language
large language
it introduces two key capabilities automated feedback generation using a fine-tuned large language model and visualization of student code submissions to uncover learning patterns
however due to the winner s curse-an issue where the policy optimization procedure exploits prediction errors rather than finding actual improvements-predicted performance improvements are often not substantiated by downstream
policy
policy learning
however due to the winner s curse-an issue where the policy optimization procedure exploits prediction errors rather than finding actual improvements-predicted performance improvements are often not substantiated by downstream policy optimization
2024 develops riesz regression for automatic
debiased
debiased machine
2024 develops riesz regression for automatic debiased machine learning which directly estimates the riesz representer or equivalently the bias-correction term by minimizing the mean squared error
scanning tunnelling microscopy stm enables atomic-resolution
imaging
electron microscopy
scanning tunnelling microscopy stm enables atomic-resolution imaging and atom manipulation but its utility is often limited by tip degradation and slow serial data acquisition
people tend to walk in groups and interactions with those
groups
human mobility
people tend to walk in groups and interactions with those groups have a significant impact on crowd behavior and pedestrian traffic dynamics
this study provides evidence-based strategies for culturally responsive
mental
mental health
this study provides evidence-based strategies for culturally responsive mental health interventions addressing persistent disparities in black men s service utilisation
comparative studies have been done by placing the image plane in five different positions on the propagation path of the
beam
beam shaping
comparative studies have been done by placing the image plane in five different positions on the propagation path of the beam so that we can detect collimated diverging and converging outputs
we demonstrate that our method achieves higher success rates on complex contact-rich tasks than end-to-end
rl
motion planning
we demonstrate that our method achieves higher success rates on complex contact-rich tasks than end-to-end rl approaches and produces more efficient coordinated behaviors than traditional sequential-only planners
we contribute design recommendations for creating inclusive mobile fitness tools that align with older
adults
older adults
we contribute design recommendations for creating inclusive mobile fitness tools that align with older adults routines and capabilities offering insights for future long-term real-world deployments
convergence analysis for an implementable scheme to solve the linear-quadratic stochastic optimal control problem with
stochastic
quadratic programming
convergence analysis for an implementable scheme to solve the linear-quadratic stochastic optimal control problem with stochastic wave equation
this could hopefully represent a first step into better insights into
viral
viral replication
this could hopefully represent a first step into better insights into viral dynamics that may help clinicians to achieve consistently better outcomes
our findings provide actionable insights into how values are learned during post-training and help to inform data curation as well as the selection of models and algorithms for
preference
preference optimization
our findings provide actionable insights into how values are learned during post-training and help to inform data curation as well as the selection of models and algorithms for preference optimization to improve model alignment to human values
to formalize interventional constraints we propose a metric to quantify total causal effects for linear
causal
interventional constraints
to formalize interventional constraints we propose a metric to quantify total causal effects for linear causal models and formulate the problem as a constrained optimization task solved using a two-stage constrained optimization method
exploring emergent topological properties in socio-economic networks through
learning
emergent behaviors
exploring emergent topological properties in socio-economic networks through learning heterogeneity
this enables dynamic and real-time tasks that were previously believed to be unattainable by large
vla
vla models
this enables dynamic and real-time tasks that were previously believed to be unattainable by large vla models
furthermore providing faithful explanations for predicted
emotions
empathic prompting
furthermore providing faithful explanations for predicted emotions is crucial to ensure interpretability and build user trust
for this problem we develop a direct debiased
machine
debiased machine
for this problem we develop a direct debiased machine learning framework with an end-to-end algorithm
we then show that electron-phonon interactions critically
modify
optical properties
we then show that electron-phonon interactions critically modify optical spectra and exciton lifetimes at finite temperatures
bridging prediction and attribution identifying forward and backward causal influence ranges using assimilative
causal
causal effects
bridging prediction and attribution identifying forward and backward causal influence ranges using assimilative causal inference
we also prove for the first time the non-singularity of the
gradient
gradient flow
we also prove for the first time the non-singularity of the gradient descent gd map on the loss landscape of realistic neural network architectures with fully connected convolutional or softmax attention layers and piecewise analytic activations which includes sigmoid relu leaky relu etc
we propose to combine the reconstruction loss with training for dynamic
correspondence
pose estimation
we propose to combine the reconstruction loss with training for dynamic correspondence along with a visibility head and fine-tuning mast3r for point tracking using a relatively small amount of synthetic data
our results underline the potential of generative
models
quantum error correction
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
applications to the get-out-the-vote dataset and criteo uplift modeling dataset demonstrate that our method outperforms fully
nonparametric
nonparametric identification
applications to the get-out-the-vote dataset and criteo uplift modeling dataset demonstrate that our method outperforms fully nonparametric machine learning methods in identifying individuals with higher treatment effects
in each transmission block vae encodes the received echo into a latent representation that
conditions
channel state information
in each transmission block vae encodes the received echo into a latent representation that conditions ddpm to predict future target states which are then used for codebook beam selection
departing from conventional euclidean parameter spaces the ndm re-conceptualizes a
neural
neural network
departing from conventional euclidean parameter spaces the ndm re-conceptualizes a neural network as a differentiable manifold where each layer functions as a local coordinate chart and the network parameters directly parameterize a riemannian metric tensor at every point
our methods and findings offer not only practical applications for quantum networks but also lead to a deeper understanding of
multipartite
entanglement entropy
our methods and findings offer not only practical applications for quantum networks but also lead to a deeper understanding of multipartite entanglement structures
to address this we propose spoken-passage multiple-choice question
answering
question answering
to address this we propose spoken-passage multiple-choice question answering a novel subjective approach evaluating the accuracy of key information in synthesized speech and release sp-mcqa-eval an 8
we propose the first known computationally tractable
algorithm
learning algorithm
we propose the first known computationally tractable algorithm for computing the solution to the graves-lai optimization problem which in turn enables the implementation of asymptotically optimal algorithms for this bandit problem
we show that this version is np-hard even when both structures the food web and the
phylogenetic
phylogenetic diversity
we show that this version is np-hard even when both structures the food web and the phylogenetic tree are stars
understanding the diversity of star formation
histories
massive stars
understanding the diversity of star formation histories sfhs of galaxies is key to reconstructing their evolutionary paths
the cross dual encoder network comprises four essential components a global encoder a local
encoder
deep network
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
to address this we propose a one problem multiple solutions 1pns training paradigm that exposes the model to a variety of
valid
mathematical reasoning
to address this we propose a one problem multiple solutions 1pns training paradigm that exposes the model to a variety of valid reasoning trajectories and thus increases inference diversity
using photoionisation models from the photodissociation region toolbox we quantify the ism
conditions
star-forming region
using photoionisation models from the photodissociation region toolbox we quantify the ism conditions in the different regions determining that the southern star-forming regions have a high density n_h sim 10 5 cm -3 and are irradiated by a moderate uv radiation field g_0 sim 10 3 habing
a fine-grained account of functional selectivity in the cortex is
essential
higher-order visual
a fine-grained account of functional selectivity in the cortex is essential for understanding how visual information is processed and represented in the brain
emotion-coherent reasoning for multimodal
llms
models llms
emotion-coherent reasoning for multimodal llms via emotional rationale verifier
spg-cdenet spatial prior-guided cross dual encoder network for
multi-organ
multi-organ segmentation
spg-cdenet spatial prior-guided cross dual encoder network for multi-organ segmentation
reinforcement learning rl fine-tuning of large language models llms often suffers from instability due to the numerical mismatch between the
training
continual learning
reinforcement learning rl fine-tuning of large language models llms often suffers from instability due to the numerical mismatch between the training and inference policies
no statistically significant differences were observed between top-3 full-9 and
consensus
findings highlight
no statistically significant differences were observed between top-3 full-9 and consensus p 0
these paradoxes highlight the tension between quantum
theory
quantum coherence
these paradoxes highlight the tension between quantum theory and our intuitions about reality being observer-independent
we quantify the gas flows from a scale of up to several parsecs down to the sub-parsec
scale
stellar mass function
we quantify the gas flows from a scale of up to several parsecs down to the sub-parsec scale along filamentary structures in the three high-mass star-forming regions g75
simulating and experimenting with social media mobilization using
llm
large language models llms
simulating and experimenting with social media mobilization using llm agents
the ability to continually learn retain and deploy skills to accomplish goals is a key feature of
intelligent
artificial intelligence
the ability to continually learn retain and deploy skills to accomplish goals is a key feature of intelligent and efficient behavior
these findings demonstrate the potential of ai agents to move beyond process partners toward colleagues that share intent strengthen group dynamics and collaborate with
humans
artificial intelligence
these findings demonstrate the potential of ai agents to move beyond process partners toward colleagues that share intent strengthen group dynamics and collaborate with humans to advance ideas
sensory representation is typically understood through a hierarchical-causal framework where progressively
abstract
artificial neural
sensory representation is typically understood through a hierarchical-causal framework where progressively abstract features are extracted sequentially
for each operation the inputs and outputs are changes in phase relative to a
fixed
phase transitions
for each operation the inputs and outputs are changes in phase relative to a fixed bias point
from linear to nonlinear provable weak-to-strong
generalization
gradient descent
from linear to nonlinear provable weak-to-strong generalization through feature learning
we also demonstrate that the inter-layer magnetic coupling in these materials can be tuned by
strain
electronic structure
we also demonstrate that the inter-layer magnetic coupling in these materials can be tuned by strain enabling the switching between the ahe and the axionic states
however such large fixed context lengths may lead to limited
exploration
context length
however such large fixed context lengths may lead to limited exploration efficiency and redundant information
robotic assistant completing collaborative tasks with dexterous
vision-language-action
vision-language-action vla
robotic assistant completing collaborative tasks with dexterous vision-language-action models
instead the model must have somehow synthesized its own
geometry
world models
instead the model must have somehow synthesized its own geometry of atomic facts encoding global relationships between all entities including non-co-occurring ones
here we aim to characterize misconceptions that users of conversational llm-based assistants may have in
programming
llm responses
here we aim to characterize misconceptions that users of conversational llm-based assistants may have in programming contexts
for temperature spans of 5-10 k we predict a coefficient of
performance
heat conduction
for temperature spans of 5-10 k we predict a coefficient of performance of 21-40
instrumental variable methods are fundamental to causal inference when
treatment
treatment effect
instrumental variable methods are fundamental to causal inference when treatment assignment is confounded by unobserved variables
occasional oddball-like sensorimotor disruptions introduced premature
feedback
sensorimotor disruptions
occasional oddball-like sensorimotor disruptions introduced premature feedback to elicit prediction errors
reinforcement learning rl can elicit strong reasoning in large language models llms yet most open efforts focus on
math
reasoning tasks
reinforcement learning rl can elicit strong reasoning in large language models llms yet most open efforts focus on math and code
our solutions assume a constant star formation efficiency a constant mass-loading factor and that the yields are linearly dependent on the
interstellar
interstellar medium
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
the results expose limitations in current llms capacity for representation alignment highlighting the need for further research on robust alignment between
language
vision-language models vlms
the results expose limitations in current llms capacity for representation alignment highlighting the need for further research on robust alignment between language and internal agent representations
while a multi-agent approach based on large
language
large language
while a multi-agent approach based on large language models llms represents a promising strategy to surpass the capabilities of single models its success is critically dependent on synergistic team composition
keywords co-ranking sep partitioning sep merge-free algorithms sep index-space optimization sep selection and merging sep
data
data structure
keywords co-ranking sep partitioning sep merge-free algorithms sep index-space optimization sep selection and merging sep data structures
under emph heterogeneous wireless scenarios however such
designs
wireless networks
under emph heterogeneous wireless scenarios however such designs are constrained by the weakest device and inflate the update variance
the evaluation framework provides a reusable and transparent
benchmark
evaluation metrics
the evaluation framework provides a reusable and transparent benchmark for assessing clinical question answering systems where semantic precision and computational efficiency are critical
self-localization on a 3d map by using an inexpensive monocular camera is required to realize
autonomous
autonomous driving
self-localization on a 3d map by using an inexpensive monocular camera is required to realize autonomous driving
beyond standard validation mobilitygen yields insights not attainable with earlier models including how access to urban space varies across
travel
travel information
beyond standard validation mobilitygen yields insights not attainable with earlier models including how access to urban space varies across travel modes and how co-presence dynamics shape social exposure and segregation
we present the first dynamic algorithms for dyck and
tree
tree edit
we present the first dynamic algorithms for dyck and tree edit distances with subpolynomial update times
using a masked autoencoding objective we pretrain reve on over 60 000 hours of eeg data from 92 datasets spanning 25 000 subjects representing the largest
eeg
electroencephalography eeg
using a masked autoencoding objective we pretrain reve on over 60 000 hours of eeg data from 92 datasets spanning 25 000 subjects representing the largest eeg pretraining effort to date
the tri-infrastructure methodology and 79-problem benchmark enable longitudinal tracking of reasoning capabilities as
foundation
foundation models
the tri-infrastructure methodology and 79-problem benchmark enable longitudinal tracking of reasoning capabilities as foundation models evolve
we propose a kinodynamic tamp framework based on a hybrid state tree that uniformly represents symbolic and numeric states during
planning
motion planning
we propose a kinodynamic tamp framework based on a hybrid state tree that uniformly represents symbolic and numeric states during planning enabling task and motion decisions to be jointly decided
furthermore for specific pairs of models and
riesz
riesz representer
furthermore for specific pairs of models and riesz representer estimation methods we can automatically obtain the covariate balancing property without explicitly solving the covariate balancing objective
we demonstrate that nonlinear quantum scrambling facilitates the achievement of super-heisenberg
scaling
quantum key distribution
we demonstrate that nonlinear quantum scrambling facilitates the achievement of super-heisenberg scaling t - beta when the generator of the parameter is time-independent
a central challenge in cognitive neuroscience is to explain how semantic and
episodic
working memory
a central challenge in cognitive neuroscience is to explain how semantic and episodic memory two major forms of declarative memory typically associated with cortical and hippocampal processing interact to support learning recall and imagination
this study bridges the gap between the two approaches by
showing
computational cost
this study bridges the gap between the two approaches by showing that both are based on essentially the same optimization problem
by first establishing a converse issf-bf theorem we reveal the equivalence among the achievability of issf by feedback the achievability of inverse optimality and the solvability of a hamilton-jacobi-isaacs equation associated with the
inverse
inverse optimal
by first establishing a converse issf-bf theorem we reveal the equivalence among the achievability of issf by feedback the achievability of inverse optimality and the solvability of a hamilton-jacobi-isaacs equation associated with the inverse optimal issf gain assignment