prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
inference latency stands as a critical bottleneck in the large-scale deployment of large
language
language models
inference latency stands as a critical bottleneck in the large-scale deployment of large language models llms
see4d pose-free 4d generation via auto-regressive
video
video generation
see4d pose-free 4d generation via auto-regressive video inpainting
the second part learns the neural stability
descriptor
neural network
the second part learns the neural stability descriptor by iteratively training the nns with sample augmentation guided by the tailored conservativeness-aware loss function
quantifying memory effects by the trace-distance based measure for non-markovianity and performing numerical simulations employing the hierarchical equations of motion approach we find a pronounced peak structure when plotting the non-markovianity
measure
memory effects
quantifying memory effects by the trace-distance based measure for non-markovianity and performing numerical simulations employing the hierarchical equations of motion approach we find a pronounced peak structure when plotting the non-markovianity measure as a function of the driving amplitude
through sed fitting we determine the luminosities of these
quasars
dwarf galaxies
through sed fitting we determine the luminosities of these quasars and find that their dust torus sizes follow the established r_ dust -l_ agn relation reported in previous studies
second we propose vimogen a flow-matching-based
diffusion
diffusion models
second we propose vimogen a flow-matching-based diffusion transformer that unifies priors from mocap data and vigen models through gated multimodal conditioning
on the go with ar attention to virtual and
physical
mobile ar
on the go with ar attention to virtual and physical targets while varying augmentation density
we study a family of stationary hamilton-jacobi-bellman hjb equations in hilbert spaces arising from stochastic optimal
control
optimal control
we study a family of stationary hamilton-jacobi-bellman hjb equations in hilbert spaces arising from stochastic optimal control problems
the optimal control model supports the development and seasonal timing of cost-effective mosquito
control
optimal control
the optimal control model supports the development and seasonal timing of cost-effective mosquito control methods
the results revealed that pedestrian speed has a negative relationship with
crash
crash risk
the results revealed that pedestrian speed has a negative relationship with crash risk while flow and speed of motorized vehicles pedestrian flow and non-motorized vehicles conflicting speed contribute positively
inference latency stands as a critical bottleneck in the large-scale deployment of
large
models llms
inference latency stands as a critical bottleneck in the large-scale deployment of large language models llms
computer-using agents powered by vision-language
models
ai systems
computer-using agents powered by vision-language models vlms have demonstrated human-like capabilities in operating digital environments like mobile platforms
our methodology is twofold first we develop algorithms to process raw position data accurately extracting the commuting trajectory
transportation
traffic dynamics
our methodology is twofold first we develop algorithms to process raw position data accurately extracting the commuting trajectory transportation mode and transfer stations
we propose viz-coast a method of leveraging the common-sense spatial reasoning of large pretrained vision-language
models
vision-language models vlms
we propose viz-coast a method of leveraging the common-sense spatial reasoning of large pretrained vision-language models to identify issues with downward refinement a priori bypassing the need to fix these failures during planning
wave guides for classical electromagnetic fields can realize the
quantum
quantum emitters
wave guides for classical electromagnetic fields can realize the quantum evolution of the wave function for a system of qubits
we find that the extracted bandgap of gaas is approximately 13 mev higher than the expected value based on
previous
bulk gaas
we find that the extracted bandgap of gaas is approximately 13 mev higher than the expected value based on previous absorbance measurements
violations of conditional exchangeability substantially limit the validity of ite estimates from
causal
conditional exchangeability
violations of conditional exchangeability substantially limit the validity of ite estimates from causal ml models in routinely collected observational data
to advance empirical practice we propose a simple joint inference procedure for the rd effect and its local external
validity
external validity
to advance empirical practice we propose a simple joint inference procedure for the rd effect and its local external validity building on calonico cattaneo and titiunik 2014 econometrica and dong and lewbel 2015 review of economics and statistics
we define symmetric and asymmetric branching
trees
phylogenetic tree
we define symmetric and asymmetric branching trees a class of processes particularly suited for modeling genealogies of inhomogeneous populations where individuals may reproduce throughout life
we propose a bias correction procedure to mitigate the adverse effects of
synthetic
synthetic data
we propose a bias correction procedure to mitigate the adverse effects of synthetic data enhancing prediction accuracy while avoiding overfitting
it not only provides a more quality foundational network framework for network research but also serve as the brand
new
scale-free networks
it not only provides a more quality foundational network framework for network research but also serve as the brand new paradigm for bridging the conceptual divide between various classical network models
large language models llms are widely used in generative
applications
llm inference
large language models llms are widely used in generative applications such as chatting code generation and reasoning
at each time step each individual produces a large number of offspring that inherit the
fitness
population genetics
at each time step each individual produces a large number of offspring that inherit the fitness of their parents up to independent and identically distributed fluctuations
building on the eth matrix ansatz and the structure of the out-of-time-order correlator otoc we show that the chaos bound directly constrains the
error
quantum correlations
building on the eth matrix ansatz and the structure of the out-of-time-order correlator otoc we show that the chaos bound directly constrains the error of an approximate quantum error-correcting code
we review self-organization in driven amorphous solids concluding with a discussion of what self-organization in driven disordered
systems
molecular dynamics
we review self-organization in driven amorphous solids concluding with a discussion of what self-organization in driven disordered systems can teach us about how simple organisms sense and adapt to their changing environments
a common approach is to train a machine learning model to predict counterfactual outcomes and then select the
policy
policy optimization
a common approach is to train a machine learning model to predict counterfactual outcomes and then select the policy that optimizes the predicted objective value
the reasoning capabilities of large language
models
large language models llms
the reasoning capabilities of large language models llms have led to their increasing employment in several critical applications particularly education where they support problem-solving tutoring and personalized study
to overcome this the framework introduces a non-linear link function within a bayesian generalized extreme value gev structure to capture
traffic
traffic dynamics
to overcome this the framework introduces a non-linear link function within a bayesian generalized extreme value gev structure to capture traffic variability more accurately
dynamic spatial treatment effect boundaries a continuous
functional
dynamic spatial
dynamic spatial treatment effect boundaries a continuous functional framework from navier-stokes equations
we observe shallow slopes in the underlying m_ rm hi -m_ star scaling relations suggesting the presence of an upper hi mass limit beyond which a galaxy can no longer retain further
hi
hi mass
we observe shallow slopes in the underlying m_ rm hi -m_ star scaling relations suggesting the presence of an upper hi mass limit beyond which a galaxy can no longer retain further hi gas
this is a concise pedagogical introduction to the dynamic field of
open
open quantum
this is a concise pedagogical introduction to the dynamic field of open quantum systems governed by markovian master equations
since the bias-correction term h_0 is essential for
ate
ate estimation
since the bias-correction term h_0 is essential for ate estimation this direct approach is expected to improve estimation accuracy for the ate
extensive simulations and a real-world application demonstrate that our approach reliably recovers unbiased
causal
causal effects
extensive simulations and a real-world application demonstrate that our approach reliably recovers unbiased causal estimates whenever exposure and confounder signals are separable on a plurality of basis functions
finally we present a proof-of-principle simulation on the ibm quantum experience platform realizing a quantum switch of two measurement
channels
quantum correlations
finally we present a proof-of-principle simulation on the ibm quantum experience platform realizing a quantum switch of two measurement channels with tunable strengths and experimentally confirming the predicted efficiency enhancement enabled by correlation-assisted superposed causal order
our comprehensive synthetic data pipeline unifies diverse generators including
stochastic
generative models
our comprehensive synthetic data pipeline unifies diverse generators including stochastic differential equations gaussian processes and audio synthesis with novel augmentations
through a systematic survey of 258 chi papers from 2020-2025 on llms we discuss how hci hardly perceives
llm
llm reasoning
through a systematic survey of 258 chi papers from 2020-2025 on llms we discuss how hci hardly perceives llm reasoning as a product of sociotechnical orchestration and often references it as an object of application
model-independent inference of galaxy star formation
histories
bulge stars
model-independent inference of galaxy star formation histories in the local volume
accurate world models are essential for enabling agents to think
plan
world models
accurate world models are essential for enabling agents to think plan and reason effectively in complex dynamic settings
to address this we propose a novel approach the
emotional
empathic prompting
to address this we propose a novel approach the emotional rationale verifier erv and an explanation reward
rectified flow models have become a de facto standard in image
generation
image generation
rectified flow models have become a de facto standard in image generation due to their stable sampling trajectories and high-fidelity outputs
conditional forecasts and proper scoring rules for reliable and
accurate
proper scoring
conditional forecasts and proper scoring rules for reliable and accurate performative predictions
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
optimal control
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
lifting and partial smoothing for stationary hjb equations and related
control
optimal control
lifting and partial smoothing for stationary hjb equations and related control problems in infinite dimensions
experiments on real world gui tasks further validate the close link between
gui
gui knowledge
experiments on real world gui tasks further validate the close link between gui knowledge and task success
while large language models llms have demonstrated remarkable performance across various reasoning tasks their ability to handle normative reasoning
remains
language agents
while large language models llms have demonstrated remarkable performance across various reasoning tasks their ability to handle normative reasoning remains underexplored
by integrating the open-mindedness of voters the stubbornness of opinion interactions and the antagonism effect manipulated by the two parties we systematically explore the intricate interplay between top-down political campaigns and bottom-up interpersonal
opinion
opinion dynamics
by integrating the open-mindedness of voters the stubbornness of opinion interactions and the antagonism effect manipulated by the two parties we systematically explore the intricate interplay between top-down political campaigns and bottom-up interpersonal opinion dynamics unveiling their nonlinear coupling impacts on...
we revisit softmax irl and show that the population maximum-likelihood solution is characterized by a linear fixed-point equation involving the
behavior
policy learning
we revisit softmax irl and show that the population maximum-likelihood solution is characterized by a linear fixed-point equation involving the behavior policy
adapting large language models llms via reinforcement learning rl is often bottlenecked by the
generation
large language models llms
adapting large language models llms via reinforcement learning rl is often bottlenecked by the generation stage which can consume over 75 of the training time
advice-seekers showed increased retrospective confidence in genai while those who declined advice showed increased
confidence
retrospective confidence
advice-seekers showed increased retrospective confidence in genai while those who declined advice showed increased confidence in self
proximal gradient algorithms pga while foundational for
inverse
gradient flow
proximal gradient algorithms pga while foundational for inverse problems like image reconstruction often yield unstable convergence and suboptimal solutions by violating the critical non-negativity constraint
autonomous robots in orchards require real-time 3d
scene
spatial reasoning
autonomous robots in orchards require real-time 3d scene understanding despite repetitive row geometry seasonal appearance changes and wind-driven foliage motion
the pilot feedback suggests the potential of the fractalbrain to
facilitate
sensorimotor disruptions
the pilot feedback suggests the potential of the fractalbrain to facilitate mindfulness and enhance attention
as a byproduct we also provide a separation between incremental and decremental algorithms for the triangle detection problem we show a decremental algorithm with
tilde
-time algorithm
as a byproduct we also provide a separation between incremental and decremental algorithms for the triangle detection problem we show a decremental algorithm with tilde o n omega total update time while every incremental algorithm requires n 3-o 1 total update time assuming the omv hypothesis
generating pivot gray codes for spanning trees of complete
graphs
tree embedding
generating pivot gray codes for spanning trees of complete graphs in constant amortized time
applying this framework to a finite-width relu network we find that its hidden layer exhibits an
abstract
deep learning
applying this framework to a finite-width relu network we find that its hidden layer exhibits an abstract representation at all global minima of the task objective
we present properties of the entire x-ray detected blazar population including the distributions of x-ray luminosities and photon indices
multiwavelength
gamma -ray
we present properties of the entire x-ray detected blazar population including the distributions of x-ray luminosities and photon indices multiwavelength properties and the blazar log n-log s distribution
we demonstrate how the metallicity dependence of the yields can be mathematically considered as a system-dependent delay time approximately equal to the system s depletion
time
stellar population
we demonstrate how the metallicity dependence of the yields can be mathematically considered as a system-dependent delay time approximately equal to the system s depletion time that when combined with system-independent delay times arising from stellar evolutionary channels produces the separation of different systems ...
however the current llm post-training paradigm faces significant data
challenges
llm post-training
however the current llm post-training paradigm faces significant data challenges including the high costs of manual annotation and diminishing marginal returns on data scales
this body of theory can accommodate a range of social dilemmas or
games
game theory
this body of theory can accommodate a range of social dilemmas or games as well as real-world complexities such as spatial structure or behaviors conditioned on reputations
comprehensive docstrings yield modest gains of 1 to 3 in functional
accuracy
findings highlight
comprehensive docstrings yield modest gains of 1 to 3 in functional accuracy though statistical significance is rare
we first empirically demonstrate that a wide range of modern
language
language models
we first empirically demonstrate that a wide range of modern language models exhibit low-rank structure in particular matrices built from the model s logits for varying sets of prompts and responses have low approximate rank
using stieltjes transform techniques we extend these results to the independent component model deriving a fixed-point equation that characterizes the limiting
spectral
spectral density
using stieltjes transform techniques we extend these results to the independent component model deriving a fixed-point equation that characterizes the limiting spectral distribution of frac 1 2 tr sigma k_n
while recent breakthroughs in computer vision cv and artificial intelligence ai have driven remarkable progress the field still faces a critical challenge as knowledge remains fragmented across
multimodal
image fusion
while recent breakthroughs in computer vision cv and artificial intelligence ai have driven remarkable progress the field still faces a critical challenge as knowledge remains fragmented across multimodal perception contextual reasoning and cooperative intelligence
however these methods assume that only one object is provided and that it is possible with the correct grasp to perform the task they are not capable of identifying grasping and using the best object for a task when many are available especially when the optimal
tool
robotic manipulation
however these methods assume that only one object is provided and that it is possible with the correct grasp to perform the task they are not capable of identifying grasping and using the best object for a task when many are available especially when the optimal tool is absent
in this project we seek to combine the two problems into a voter-blotto game and examine what components of the graph most
effect
swing voters
in this project we seek to combine the two problems into a voter-blotto game and examine what components of the graph most effect its value in the eyes of the competing players
key to our analysis is a case-by-case investigation based on the nerve complex of the set of maximal
codewords
neural codes
key to our analysis is a case-by-case investigation based on the nerve complex of the set of maximal codewords of a neural code
a data-driven method to determine combination
averaging
weight constraints
a data-driven method to determine combination averaging weights typically optimizes a criterion under certain weight constraints
however brain signals infused with prior knowledge and associations exhibit a significant information asymmetry when compared to raw visual features still posing challenges for decoding
fmri
visual stimuli
however brain signals infused with prior knowledge and associations exhibit a significant information asymmetry when compared to raw visual features still posing challenges for decoding fmri representations under the supervision of images
validated in high-fidelity simulation with realistic quadrotor dynamics the resulting policies significantly outperform both a standard
reinforcement
deep reinforcement
validated in high-fidelity simulation with realistic quadrotor dynamics the resulting policies significantly outperform both a standard reinforcement learning baseline and a state-of-the-art game-theoretic planner
to illustrate this we construct a class of structured datasets where incremental adam provably converges to the ell_2 -max-margin classifier in contrast to the ell_ infty -max-margin
bias
debiased machine learning
to illustrate this we construct a class of structured datasets where incremental adam provably converges to the ell_2 -max-margin classifier in contrast to the ell_ infty -max-margin bias of full-batch adam
second it incorporates human-written code submissions collected from real programming contests enabling both quantitative comparisons and qualitative analyses of llm outputs against
human-written
code review
second it incorporates human-written code submissions collected from real programming contests enabling both quantitative comparisons and qualitative analyses of llm outputs against human-written codes
the goal of policy learning is to train a
policy
policy learning
the goal of policy learning is to train a policy function that recommends a treatment given covariates to maximize population welfare
adaptive trajectory refinement for optimization-based local
planning
obstacle avoidance
adaptive trajectory refinement for optimization-based local planning in narrow passages
here we present an analytically-tractable model of visual rivalry that quantitatively explains the hysteretic transition between periods of
awareness
visual rivalry
here we present an analytically-tractable model of visual rivalry that quantitatively explains the hysteretic transition between periods of awareness and suppression in tcfs
the performance of artificial intelligence ai
systems
trustworthy ai
the performance of artificial intelligence ai systems fundamentally depends on high-quality training data
we also provide new data-driven control design methods in terms of
linear
linear control
we also provide new data-driven control design methods in terms of linear matrix inequalities that complement the conditions for informativity
finally we use the connection to instanton rate
theory
instanton theory
finally we use the connection to instanton rate theory which is also derived from flux correlation functions to discuss the often misunderstood relationship between tunneling splittings and reaction rate constants
furthermore using a synthetic preference dataset that enables controlled manipulation of values we find that different preference optimization algorithms lead to different value alignment outcomes even when
preference
preference optimization
furthermore using a synthetic preference dataset that enables controlled manipulation of values we find that different preference optimization algorithms lead to different value alignment outcomes even when preference data is held constant
these results highlight the significant room for improving the mathematical
reasoning
large language models llms
these results highlight the significant room for improving the mathematical reasoning in current llms
recent progress in large language models has made them increasingly capable research
assistants
large language
recent progress in large language models has made them increasingly capable research assistants in mathematics
to overcome these challenges we then develop a generic synthesis framework based on the flow of neural dynamics drift enabling explicit piecewise constant and
constant-in-time
dynamical systems
to overcome these challenges we then develop a generic synthesis framework based on the flow of neural dynamics drift enabling explicit piecewise constant and constant-in-time inputs
our result nearly matches the o log 2 n approximation
guarantee
online algorithm
our result nearly matches the o log 2 n approximation guarantee of the quasi-polynomial-time algorithm by li xu and zhang icalp 2025
greenberger-horne-zeilinger ghz states play a central role in quantum
computing
quantum dot
greenberger-horne-zeilinger ghz states play a central role in quantum computing and communication protocols as a typical multipartite entanglement resource
the estimation method linear smoothers maximum likelihood generalized method of
moments
existing methods
the estimation method linear smoothers maximum likelihood generalized method of moments etc
we demonstrate both the generation and detection of an ultra-high flux of polarization bell
states
quantum emitters
we demonstrate both the generation and detection of an ultra-high flux of polarization bell states using broadband hyper-entangled bi-photons that are quantum-correlated in both polarization and time-energy
this article aims to deliver a comprehensive and up-to-date review of recent developments on sequential stopping rules intentionally emphasizing standard and moderately generalized monte carlo methods which have historically served and likely will continue to serve as fundamental bases for both theoretical and practica...
stopping
stopping rules
this article aims to deliver a comprehensive and up-to-date review of recent developments on sequential stopping rules intentionally emphasizing standard and moderately generalized monte carlo methods which have historically served and likely will continue to serve as fundamental bases for both theoretical and practica...
we partially answer this by showing that for every d otimes d
entangled
multipartite entanglement
we partially answer this by showing that for every d otimes d entangled state with even d there exist three projective measurements which are antidiscriminable but not discriminable with that input state but those three measurements are not antidiscriminable with the product probe
our algorithm is given in an extended form that supports
approximate
-approximation algorithm
our algorithm is given in an extended form that supports approximate preservation it minimizes a finite fuzzy interpretation mathcal i while preserving fuzzy concept assertions up to a degree gamma in 0 1
this survey bridges that gap by delivering a forward-looking analysis of object detection in avs emphasizing emerging paradigms such as
vision-language
vision-language models vlms
this survey bridges that gap by delivering a forward-looking analysis of object detection in avs emphasizing emerging paradigms such as vision-language models vlms large language models llms and generative ai rather than re-examining outdated techniques
the rotation speed of the spiral pattern in the milky
way
dwarf galaxies
the rotation speed of the spiral pattern in the milky way galaxy
we introduce the environment configuration diagnosis benchmark enconda-bench which provides process-level trajectory assessment of fine-grained agent capabilities during environment setup-planning perception-driven error diagnosis feedback-driven repair and action to execute final
environment
environment configuration
we introduce the environment configuration diagnosis benchmark enconda-bench which provides process-level trajectory assessment of fine-grained agent capabilities during environment setup-planning perception-driven error diagnosis feedback-driven repair and action to execute final environment configuration
approximately 80 of the emission of a region
originates
vi emission
approximately 80 of the emission of a region originates from the halo component
reinforcement learning rl offers a data-driven approach to derive control
policies
control strategy
reinforcement learning rl offers a data-driven approach to derive control policies for such challenges
at low weber numbers only partial rebound was observed while at
intermediate
weber numbers
at low weber numbers only partial rebound was observed while at intermediate values droplets rebounded completely
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained bilevel
optimization
optimization problem
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained bilevel optimization sflcb
to further enhance performance an adaptive active ris configuration strategy is employed which refines the
beam
beam pattern
to further enhance performance an adaptive active ris configuration strategy is employed which refines the beam direction based on an initial user location estimate
compared with the local gaussian correlation network among positive tails and the conventional pearson correlation network the properties of the local gaussian
correlation
correlation network
compared with the local gaussian correlation network among positive tails and the conventional pearson correlation network the properties of the local gaussian correlation network among negative tails are more sensitive to the stock market risks
allowing the order of quantum operations to exist in superposition is known to
open
quantum advantage
allowing the order of quantum operations to exist in superposition is known to open new routes for thermodynamic tasks
to illustrate the contrast between ei-inspired systems and traditional architectures that decouple sensing computation and actuation we present and discuss a collection of robots developed by the author and his team at the autonomous microrobotic
systems
robotic systems
to illustrate the contrast between ei-inspired systems and traditional architectures that decouple sensing computation and actuation we present and discuss a collection of robots developed by the author and his team at the autonomous microrobotic systems laboratory amsl
in ate estimation the balancing weights and the regression functions of the outcome play important roles where the
balancing
covariate balancing
in ate estimation the balancing weights and the regression functions of the outcome play important roles where the balancing weights are referred to as the riesz representer bias-correction term and clever covariates depending on the context