prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
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
waveguide modes
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
our work not only reveals the key role of anisotropic epc in controlling the thermal and optical properties of tairte4 but also provides insights into designing polarization-sensitive optoelectronic
devices
photonic circuits
our work not only reveals the key role of anisotropic epc in controlling the thermal and optical properties of tairte4 but also provides insights into designing polarization-sensitive optoelectronic devices based on topological semimetals
motivated by practical applications we use this baseline to develop a new framework for fast approximate matrix multiplication
amm
approximation factor
motivated by practical applications we use this baseline to develop a new framework for fast approximate matrix multiplication amm via low-degree approximations of the cksu polynomials
our key idea is to decouple visual and linguistic adaptation by introducing two lightweight modules a domain classifier to identify the input image type and a dual adapter mechanism comprising a prompt adapter for
language
vision-language models vlms
our key idea is to decouple visual and linguistic adaptation by introducing two lightweight modules a domain classifier to identify the input image type and a dual adapter mechanism comprising a prompt adapter for language modulation and a visual adapter for vision feature adjustment
the orbital angular momentum oam of light is a versatile degree of freedom with transformative impact across optical
communication
optical communication
the orbital angular momentum oam of light is a versatile degree of freedom with transformative impact across optical communication imaging and micromanipulation
we conducted a simulation study to examine confounding bias in ite estimates generated by
causal
causal effect
we conducted a simulation study to examine confounding bias in ite estimates generated by causal forest and x-learner models under varying conditions including the presence or absence of true heterogeneity
adapting large language models llms via reinforcement learning rl is often bottlenecked by the
generation
large language
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
we propose a functional evaluation metric for
generative
generative models
we propose a functional evaluation metric for generative models based on the relative density ratio rdr designed to characterize distributional differences between real and generated samples
self-supervised synthetic pretraining for inference of stellar mass embedded in
dense
dense gas
self-supervised synthetic pretraining for inference of stellar mass embedded in dense gas
our results provide fundamental insights into light-matter interactions in
solids
optical properties
our results provide fundamental insights into light-matter interactions in solids at the nanoscale and are vital for optimally designing the new generation of absorption-based flexible optoelectronic devices
we theoretically analyze the convergence behavior of the proposed scheme and quantify its gains in expected
communication
channel estimation
we theoretically analyze the convergence behavior of the proposed scheme and quantify its gains in expected communication efficiency and training accuracy
the price-pareto growth model of networks with
community
mobility networks
the price-pareto growth model of networks with community structure
we show that a subtle modification of standard bifurcation analysis identifies such
critical
phase transition
we show that a subtle modification of standard bifurcation analysis identifies such critical numbers including those associated with discreteness- and noise-induced transitions
to solve this high-dimensional non-convex problem under uncertain channels we develop a deep
reinforcement
reinforcement learning
to solve this high-dimensional non-convex problem under uncertain channels we develop a deep reinforcement learning solution framework based on the proximal policy optimization ppo algorithm that integrates distribution-aware action modeling and a multi-branch actor network
although the evaluation was limited to simulation these results establish predictive
processing
predictive processing
although the evaluation was limited to simulation these results establish predictive processing as a universal and scalable computational principle pointing toward robust flexible and autonomous caregiving robots while offering theoretical insight into the human brain s ability to achieve flexible adaptation in uncerta...
in this work we propose a flow decomposition-and-aggregation framework built upon an
inversion-free
flow matching
in this work we propose a flow decomposition-and-aggregation framework built upon an inversion-free formulation to address these limitations
in this work we propose cola-world which for the first time successfully realizes this synergistic paradigm resolving the core challenge in joint learning through a critical warm-up phase that effectively aligns the representations of the from-scratch lam with the pre-trained
world
world models
in this work we propose cola-world which for the first time successfully realizes this synergistic paradigm resolving the core challenge in joint learning through a critical warm-up phase that effectively aligns the representations of the from-scratch lam with the pre-trained world model
discrete dynamics arise naturally in systems with broken temporal translation
symmetry
quantum walk
discrete dynamics arise naturally in systems with broken temporal translation symmetry and are typically described by first-order recurrence relations representing classical or quantum markov chains
modulating the free-electron wave function with light brings new opportunities to create attosecond electron pulse trains to probe the
quantum
quantum dot
modulating the free-electron wave function with light brings new opportunities to create attosecond electron pulse trains to probe the quantum coherence of systems with significantly improved spatial resolution and to generate classical and non-classical states of light with wide tunability
normal curves in sub-finsler lie groups branching for strongly
convex
strongly convex
normal curves in sub-finsler lie groups branching for strongly convex norms and face stability for polyhedral norms
programming assistants powered by large language
models
large language
programming assistants powered by large language models llms have become widely available with conversational assistants like chatgpt proving particularly accessible to less experienced programmers
in this paper we show new strongly polynomial work-depth tradeoffs for computing single-source shortest paths sssp in non-negatively weighted directed
graphs
polynomial time
in this paper we show new strongly polynomial work-depth tradeoffs for computing single-source shortest paths sssp in non-negatively weighted directed graphs in parallel
we combine deep photometric data in the cosmos and xmm-lss fields with high-resolution
cosmological
host galaxy
we combine deep photometric data in the cosmos and xmm-lss fields with high-resolution cosmological hydrodynamical simulations to explore two key questions 1 how does the galaxy stellar mass function particularly in the dwarf mstar 10 9
finally type iii-d galaxies have low mass surface density
disks
circumgalactic medium
finally type iii-d galaxies have low mass surface density disks sigma delta r_ mathrm exp sim 0
in this paper we investigate this idea in the classical paradigm of the ultimatum
game
game theory
in this paper we investigate this idea in the classical paradigm of the ultimatum game which we theoretically modify to introduce prejudice at the level of players terming its intensity as prejudicity
using numerical simulations we quantify the degradation in performance due to disorder and identify single-qubit rotations two-qubit entangling gates and quantum information transport as
particularly
qubit readout
using numerical simulations we quantify the degradation in performance due to disorder and identify single-qubit rotations two-qubit entangling gates and quantum information transport as particularly susceptible
the fact that our algorithm works for typical uniformly random constant degree regular graphs rather than for all constant degree graphs is unavoidable thanks to the following impossibility result that we obtain for every fixed k in n the approximation factor of any algorithm for average distance that works for all con...
graphs
-approximation algorithm
the fact that our algorithm works for typical uniformly random constant degree regular graphs rather than for all constant degree graphs is unavoidable thanks to the following impossibility result that we obtain for every fixed k in n the approximation factor of any algorithm for average distance that works for all con...
in the data-scarce regime generalization occurs via benign overfitting or fails via harmful
overfitting
machine learning
in the data-scarce regime generalization occurs via benign overfitting or fails via harmful overfitting depending on the amount of data and we characterize the transition boundary
despite recent advances in 3d human motion
generation
video generation
despite recent advances in 3d human motion generation mogen on standard benchmarks existing models still face a fundamental bottleneck in their generalization capability
to overcome the potential non-smoothness of the hyper-objective and the computational challenges associated with the hessian matrix we utilize penalty and augmented lagrangian methods to reformulate the original
problem
minimax optimal
to overcome the potential non-smoothness of the hyper-objective and the computational challenges associated with the hessian matrix we utilize penalty and augmented lagrangian methods to reformulate the original problem as a single-level one
it introduces two key capabilities automated feedback generation using a fine-tuned large language model and visualization of student
code
code review
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
we show that if the system is controllable then incorporating this as prior knowledge does not relax the conditions required for
data-driven
data-driven stabilization
we show that if the system is controllable then incorporating this as prior knowledge does not relax the conditions required for data-driven stabilization
specifically we propose an additive instrumental variable framework to identify mean potential
outcomes
potential outcomes
specifically we propose an additive instrumental variable framework to identify mean potential outcomes and the average treatment effect with a weighting function
while these machine learning models boast impressive accuracy a related concern is how to assess and maintain
calibration
machine learning
while these machine learning models boast impressive accuracy a related concern is how to assess and maintain calibration in the predictions these models make
we design polynomial-time approximations to the optimum online algorithm achieving guarantees of 7 8 for vertex-weighted
graphs
-approximation algorithm
we design polynomial-time approximations to the optimum online algorithm achieving guarantees of 7 8 for vertex-weighted graphs and 2 sqrt 2 -2 approx 0
the results establish the first unified baseline of
photonic
single photons
the results establish the first unified baseline of photonic machine-learning performance revealing complementary strengths between variational hardware-native and hybrid approaches
this work introduces steervlm a lightweight steering module designed to guide
vision-language
vision-language models vlms
this work introduces steervlm a lightweight steering module designed to guide vision-language models vlms towards outputs that better adhere to desired instructions
by examining the challenges in data-efficient llm
post-training
training data
by examining the challenges in data-efficient llm post-training we highlight open problems and propose potential research avenues
while many studies rely on branch length information the topology of
phylogenetic
phylogenetic tree
while many studies rely on branch length information the topology of phylogenetic trees particularly their degree of imbalance offers a robust framework for inferring evolutionary dynamics when timing data is uncertain
one experiment is robustly more informative than another if the decision maker s maxmin expected utility after observing the output of the former is always at least her maxmin expected
utility
randomized experiments
one experiment is robustly more informative than another if the decision maker s maxmin expected utility after observing the output of the former is always at least her maxmin expected utility after observing the latter
our results demonstrate that roboos-next achieves superior performance across heterogeneous embodiments validating its effectiveness in enabling lifelong scalable and robust
multi-robot
robotic systems
our results demonstrate that roboos-next achieves superior performance across heterogeneous embodiments validating its effectiveness in enabling lifelong scalable and robust multi-robot collaboration
non-monotone submodular maximization subject to a matroid
constraint
submodular maximization
non-monotone submodular maximization subject to a matroid constraint under noise
a unified framework for spatial and temporal treatment effect
boundaries
treatment effect
a unified framework for spatial and temporal treatment effect boundaries theory and identification
consequently the average performance achieved by llms remains
considerably
superior performance
consequently the average performance achieved by llms remains considerably below the human baseline
recent advances in data collection and technology enable a deeper understanding of complex urban commuting yet few studies have rigorously analyzed the temporal stability and origin-destination od heterogeneity of
route
route choice
recent advances in data collection and technology enable a deeper understanding of complex urban commuting yet few studies have rigorously analyzed the temporal stability and origin-destination od heterogeneity of route choice
the tip density reaches sim 2 mathrm cm -3 implying an ambient medium density of sim 10 -3
mathrm
interstellar medium
the tip density reaches sim 2 mathrm cm -3 implying an ambient medium density of sim 10 -3 mathrm cm -3 in agreement with the galactic warm ionized medium at a distance of sim 5 kpc
such simulations -- for which classical methods are often inaccurate -- are critical to advancing our knowledge and understanding of quantum chemistry and materials underpinning a wide range of fields from
biochemistry
numerical simulations
such simulations -- for which classical methods are often inaccurate -- are critical to advancing our knowledge and understanding of quantum chemistry and materials underpinning a wide range of fields from biochemistry to clean-energy technologies and chemical synthesis
we find that dwarf agn selected by infrared colors are the most distinct
population
active galactic
we find that dwarf agn selected by infrared colors are the most distinct population with the highest star formation rates and lowest stellar masses
to mitigate the prohibitive overhead associated with full
channel
channel state information csi
to mitigate the prohibitive overhead associated with full channel state information at the transmitter csit we propose a partial-csit-based beamforming scheme that leverages randomized steering vectors and limited user-side feedback based on signal quality measurements
we show that such collections of relation decoders can be highly compressed by simple order-3 tensor networks without significant loss in
decoding
sparse autoencoders
we show that such collections of relation decoders can be highly compressed by simple order-3 tensor networks without significant loss in decoding accuracy
we investigate whether large language models
llms
language models
we investigate whether large language models llms can act as in-context meta-learners for this task
the focus is on the mathematical description of these interactions and their role in deriving differential
systems
complex systems
the focus is on the mathematical description of these interactions and their role in deriving differential systems that describe the aforementioned dynamics
these results highlight the significant room for improving the mathematical
reasoning
reasoning curriculum
these results highlight the significant room for improving the mathematical reasoning in current llms
firstly our deep learning model predicts correspondence probabilities and reliabilities for every pair of a
trajectory
autonomous driving
firstly our deep learning model predicts correspondence probabilities and reliabilities for every pair of a trajectory and sensor measurements
the evolutionary mechanisms of cooperative behavior represent a fundamental topic in complex systems and
evolutionary
game theory
the evolutionary mechanisms of cooperative behavior represent a fundamental topic in complex systems and evolutionary dynamics
for bounded treewidth permutation classes which include the above-mentioned separable class we further reduce the
space
tree embedding
for bounded treewidth permutation classes which include the above-mentioned separable class we further reduce the space overhead to a lower order additive term making our data structure succinct
experimenting with llama-3 and qwen-3 models of different sizes and popular supervised fine-tuning sft and preference optimization datasets and algorithms we find that the sft phase generally establishes a model s values and subsequent
preference
preference learning
experimenting with llama-3 and qwen-3 models of different sizes and popular supervised fine-tuning sft and preference optimization datasets and algorithms we find that the sft phase generally establishes a model s values and subsequent preference optimization rarely re-aligns these values
by varying the parameters within the objective function and the constraints we determine how the optimal
spatial
spatial structure
by varying the parameters within the objective function and the constraints we determine how the optimal spatial structure may vary when individuals differ in their information gathering ability and how this variation differs in the context of resource constraints
by moving from monolithic models to orchestrated intelligence this approach seeks to align medical
ai
artificial intelligence
by moving from monolithic models to orchestrated intelligence this approach seeks to align medical ai with the first principle of medicine care that is transparent equitable and centered on the individual
large language models llms are increasingly used as
raters
llm raters
large language models llms are increasingly used as raters for evaluation tasks
to test this we conducted a randomized controlled experiment n 486 comparing a two variants of reflective human-led modes in which the llm elicits elaboration through suggestions or questions against b a proactive model-led mode in which the
llm
llm responses
to test this we conducted a randomized controlled experiment n 486 comparing a two variants of reflective human-led modes in which the llm elicits elaboration through suggestions or questions against b a proactive model-led mode in which the llm independently rewrites ideas
the coordination game payoff structure captures the insight that mutualistic
strategies
control strategies
the coordination game payoff structure captures the insight that mutualistic strategies lead to robust advantages only after such biological markets reach a certain scale
our framework unifies riesz regression for automatic
debiased
debiased machine learning
our framework unifies riesz regression for automatic debiased machine learning covariate balancing targeted maximum likelihood estimation tmle and density-ratio estimation
a view-conditional video inpainting model is trained to learn a robust geometry prior by denoising realistically synthesized warped
images
computer vision
a view-conditional video inpainting model is trained to learn a robust geometry prior by denoising realistically synthesized warped images and to inpaint occluded or missing regions across virtual viewpoints eliminating the need for explicit 3d annotations
high-resolution x-ray data are best suited for outflow s studies and the observed
absorption
ionized gas
high-resolution x-ray data are best suited for outflow s studies and the observed absorption lines on heavy elements are evidence of the physical properties of an absorbing gas
along the way we also develop improved indexes for short patterns offering
better
pattern matching
along the way we also develop improved indexes for short patterns offering better trade-offs in this practically relevant special case
these results demonstrate that spiral density waves can persist in fully cosmological
disks
circumgalactic medium
these results demonstrate that spiral density waves can persist in fully cosmological disks linking internal dynamical processes to galaxy assembly and offering testable predictions for present and future surveys such as jwst and roman
we conduct a comprehensive comparison between redllm pretrained with prefix language
modeling
large language models llms
we conduct a comprehensive comparison between redllm pretrained with prefix language modeling lm and decllm pretrained with causal lm at different model scales ranging from sim 150m to sim 8b
we further establish a strong baseline solution that outperforms prior
approaches
existing methods
we further establish a strong baseline solution that outperforms prior approaches under these challenging conditions
while large language models have been applied to energy systems as code generators and parameter extractors no existing implementation deploys llms as autonomous coordinators managing the complete workflow from natural
language
language models
while large language models have been applied to energy systems as code generators and parameter extractors no existing implementation deploys llms as autonomous coordinators managing the complete workflow from natural language input to multi-appliance scheduling
while a multi-agent approach based on large language models llms represents a promising strategy to surpass the
capabilities
large language models llms
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
however the fundamental question of emph which problems with a deterministic
complexity
time complexity
however the fundamental question of emph which problems with a deterministic complexity of omega log n can be solved exponentially faster using randomization still remains wide open
we release the code and data for aot-psyphybench to encourage further progress in the physical and temporal
reasoning
models vlms
we release the code and data for aot-psyphybench to encourage further progress in the physical and temporal reasoning capabilities of vlms
to ensure reward fidelity our automated grader calibration pipeline systematically purges noise from the llm-based
reward
human feedback
to ensure reward fidelity our automated grader calibration pipeline systematically purges noise from the llm-based reward model with minimal human supervision
to this end we introduce a novel metric for comparing both intrinsic
recurrent
dynamical systems
to this end we introduce a novel metric for comparing both intrinsic recurrent and input-driven dynamics called inputdsa idsa
while deep learning dominates recent mtl research support vector machines svms and twin
svms
machine learning
while deep learning dominates recent mtl research support vector machines svms and twin svms twsvms remain relevant due to their interpretability theoretical rigor and effectiveness with small datasets
1 pc the magnetic field lines appear roughly perpendicular to the
filament
magnetic field
1 pc the magnetic field lines appear roughly perpendicular to the filament s long axis in contrast to the smaller-scale structure sim 0
metacognition and confidence dynamics in advice taking from
generative
ai use
metacognition and confidence dynamics in advice taking from generative ai
recent work has shown that different large language models llms converge to similar and accurate
input
vision-language models
recent work has shown that different large language models llms converge to similar and accurate input embedding representations for numbers
in the limited cases where ground truth is available through exact classical
simulation
quantum mechanics
in the limited cases where ground truth is available through exact classical simulation we find that it agrees with the results we obtain from the quantum device
however the performance of all previous digital quantum simulations has been matched by classical methods and it has thus far remained unclear whether near-term intermediate-scale quantum hardware could offer any computational
advantage
quantum batteries
however the performance of all previous digital quantum simulations has been matched by classical methods and it has thus far remained unclear whether near-term intermediate-scale quantum hardware could offer any computational advantage in this area
experimental results on multiple benchmark
datasets
real-world datasets
experimental results on multiple benchmark datasets demonstrate the superior performance of the proposed method in terms of both accuracy and sparsity
we study integer programs where the constraint matrix a has such a path-like
structure
integer programs
we study integer programs where the constraint matrix a has such a path-like structure every non-zero coefficient appears in at most two consecutive constraints
this problem is inherently challenging due to its
non-convex
optimization problem
this problem is inherently challenging due to its non-convex nature
anti gravitron a statistical physics perspective on multidimensional metrics of
polarizing
gromov-wasserstein distance
anti gravitron a statistical physics perspective on multidimensional metrics of polarizing inequality
notably the phase offsets suggest structurally distinct causes of rural and urban accident
risk
crash risk
notably the phase offsets suggest structurally distinct causes of rural and urban accident risk with urban regions exhibiting increasing acceleration in accident scaling potentially linked to growth in vehicle numbers size and weight
the key task of machine learning is to minimize the loss function that measures the model fit to the
training
deep learning
the key task of machine learning is to minimize the loss function that measures the model fit to the training data
we introduce a general diploid population model with self-fertilization and possible overlapping generations and study the genealogy of a sample of n genes as the population
size
population size
we introduce a general diploid population model with self-fertilization and possible overlapping generations and study the genealogy of a sample of n genes as the population size n tends to infinity
our primary contribution is the novel imposition of explicit constraints directly within the
flow
optical flow
our primary contribution is the novel imposition of explicit constraints directly within the flow matching process ensuring that the generated trajectories adhere to vital safety and kinematic rules
current non-invasive neuroimaging techniques trade off between spatial
resolution
temporal resolution
current non-invasive neuroimaging techniques trade off between spatial resolution and temporal resolution
instrumental variable methods are fundamental to
causal
treatment effect
instrumental variable methods are fundamental to causal inference when treatment assignment is confounded by unobserved variables
interpreting visual observations and natural
language
vision-language-action vla
interpreting visual observations and natural language instructions for complex task execution remains a key challenge in robotics and ai
super-heisenberg scaling which scales as n - beta with beta 1 in terms of the number of particles n or t - beta in terms of the evolution time t is better than heisenberg
scaling
super-heisenberg scaling
super-heisenberg scaling which scales as n - beta with beta 1 in terms of the number of particles n or t - beta in terms of the evolution time t is better than heisenberg scaling in quantum metrology
skeb provides a foundation for assessing unlearning completeness
robustness
llm raters
skeb provides a foundation for assessing unlearning completeness robustness and overall behavior in llms
however even under this condition classical
proper
proper scoring
however even under this condition classical proper scoring rules fail to elicit correct forecasts
deep networks have shown remarkable performance across a wide range of tasks yet getting a global concept-level
understanding
convolutional neural
deep networks have shown remarkable performance across a wide range of tasks yet getting a global concept-level understanding of how they function remains a key challenge
data selection is a critical aspect of reinforcement learning with verifiable rewards rlvr for enhancing the
reasoning
reinforcement learning
data selection is a critical aspect of reinforcement learning with verifiable rewards rlvr for enhancing the reasoning capabilities of large language models llms
in thisproject we have used machine learning techniques like logistic regression random forest and
support
machine learning
in thisproject we have used machine learning techniques like logistic regression random forest and support vector machines to analyze the health claims data and identify demographic and medical factors that play a crucial role in predicting all-cause readmissions
this energy is believed to impact the star formation activity and contribute to the
quenching
massive galaxies
this energy is believed to impact the star formation activity and contribute to the quenching of galaxies
as a consequence the two approaches are interchangeable in several respects and share the same theoretical
guarantees
theoretical guarantees
as a consequence the two approaches are interchangeable in several respects and share the same theoretical guarantees under common conditions