text stringlengths 54 548k | label stringclasses 4
values | id_ stringlengths 32 32 |
|---|---|---|
A final application of {{formula:e54bbb68-49e0-4657-bb35-92502b2e1c6b}} -graph calculus for Legendrian weaves develops the connection of symplectic topology to V. Fock and A. Goncharov's cluster varieties of framed local systems {{cite:383b54b26f3ee8a896892710cffdd39f9d9e654c}} (see also {{cite:5d0f6f3d2877584412c88a78... | r | bd12b331b23dd178c075ab117f4c91ea |
In this work, the perceptual network {{formula:e1c356fa-6a5d-46ee-91cb-c4a9b6632608}} is a pretrained VGG16 net {{cite:25b36db187bbd4cff87e0f8337cf78a9b806a699}} and we empirically find that {{formula:009564ed-fa65-486c-ab13-34378157c274}} works well.
| m | f8557ed6f6d0d25b32ceb65f539fdb5e |
Stability. The RTI implies an obvious explanation for the need for IMP for stability: only when the network is stable enough to retrain to a similar region can we find lottery tickets. Our results in section confirm this by looking at the large-batch training on common vision tasks. This reinforces the hypothesis of {... | d | 4ae2223ccef5d52eae946716340b5d8e |
We also compare with several popular zero-shot KD methods, including
Meta-KD {{cite:d8bb79a804af9a7d79ef01d8cee3f2e221c7896d}}, ZSKD {{cite:726e88d9cf610dd477923e2f905808f11c5f5038}},
DAFL {{cite:d007ad1110db641161f2ae964c44140502600320}}, DFKD {{cite:6d32333f39b8216c251a2a38b83d67706e549a04}},
and ZSDB3KD {{cite:5e344... | m | 66313040837010c7c6e6e58416d794da |
A thorough research framework is established in {{cite:b98b1061f30951b22b70d12ffedc06be5baaaaee}} pointing out principles concerning deep neural network theories. Tacit knowledge represented by user characteristics may be quantified with mastering level of these principles in {{cite:b98b1061f30951b22b70d12ffedc06be5baa... | d | e37f8dc5366dad73dd05f8335e803664 |
According to the experiment results, MACRL shows better performance across different benchmarks. Moreover, we show that the learned representations from MACRL can be tuned more easily and efficiently (e.g., better accuracy within less epochs). Furthermore, we visualize the attention from the pre-trained weights using t... | d | a70bf97a5b6899958d7fcae06e28b339 |
Golub and Van Loan {{cite:59189020d85cef97bcadbd23a9f3e69f88ada3fd}} took errors of {{formula:0d26f3ec-dd5b-4181-a1d1-41e5659abce2}} into consideration and established the total least squares for matrix linear equations. It is aimed to solve the following problem,
{{formula:788db429-e319-44bd-8630-876bd4b0700b}}
| m | 320957d14bb17014257ae2c903b4bafe |
The attempts to detect early warning signs for critical transitions
are based on the concept of a deteriorating environment as embodied
in a changing parameter {{cite:1d65f749a2dd0f4afdb317ed2530d97f378f8a46}}, which is a different kind
of transition than one which is driven instead by stochasticity in an
environment w... | d | cb24edf3721c1294f5f365e4e7bbcf08 |
The ability to continue operating in an environment where partial failures are to be expected is a core concern in distributed computing. Checkpoint and Rollback Recovery (CPR) is the most widely used fault tolerance strategy for DSP systems. It improves the overall reliability of streaming jobs executing within them a... | i | 37937c7006de105b08c04d8bf9c2b0bf |
Corollary G.2 (TPM, from {{cite:0ab8b11e0f58eb7f6fd6d2998a7e126af2366202}})
Let {{formula:40dd34a2-6cef-4593-b4cf-d239c48e6247}} be a constant and {{formula:54509e67-4e79-493e-b577-9f9da802ed75}} and {{formula:a0f0de3f-769a-47c3-bef4-2a49d02fce22}} . Let {{formula:1e7e0f37-1580-4fca-b913-017883764e3a}} be two posit... | m | 5e49f978cbec8bf59f3e54d0f9abb815 |
Let's list three equation we are going to use {{cite:a0ddb352ac56ac2c14f7d39ae5ef5236cf949957}}, {{cite:ab73ac2a3cb4a5710087673f8725580c5e3394a9}}, {{cite:adbb027dfe02226bf74ce192819a0a77c7e5ad8d}}
{{formula:b89a720b-89fd-4f8f-bd9d-603d803575bc}}
| r | e45b0c4c003cc200b7ff7fc4c3ec16ae |
We used two methods to train our models on word embeddings, Fasttext and N-Grams {{cite:6ea61fe418cabd80b5abe77adc9d7c06023b845a}}. Fig. REF shows the working of modified N-Gram model used in this research. The n-gram model converts the document into tokens and stores these tokens in a dictionary based on the co-occur... | m | fd302720d49443285a7ebb767414a702 |
In quenched QCD at high temperatures the spectral density of eigenstates
(of the staggered Dirac operator) shows a square root behaviour
{{cite:a143bd19b684d7dbeb10a905d036ffa6be96bb73}}, and the delocalization energy scale has been
empirically identified using the inverse participation ratio
(IPR) {{cite:c387d6f73fee... | i | 1e685a7cde1c1629fb959a0771ea3ffb |
The first round (Round 0: Key Exchange) is similar to other protocols such as {{cite:a140589bde377e15b5ce236af049ac93ca7d29d4}} and {{cite:a002acfed4ca3ad4f8486fd354b5e5e527fdb0c8}},
and only involves standard PKI and symmetric key encryption primitives. This round does not need
to be executed each time a new secure
ag... | m | 136332d3adb109be5719ac0c936b4765 |
CIFAR-10 and CIFAR-100:
The MEST accuracy results are shown in Table REF . We include the results at sparsity ratios of 90%, 95%, and 98% with unstructured sparsity scheme.
Methods that use dynamic sparse training (DeepR, SET, and DSR) achieve slightly better results compared to fixed-mask sparse training.
Compared to ... | r | da57c33ee0489b8393fee710a418b2cb |
Settings To demonstrate that the proposed SLIM is able to disentangle the high level semantic attributes of featres at different scales, we conduct the style mixing experiment as done in the FastGAN paper {{cite:9420a8de32caf92929b3d0f7cc24239333518c0d}}, in which for a pair of style and content images, we extract chan... | r | b4592dbd806f5bf7f338ce72ff1e9bf5 |
Every Banach space is a Banach space of type 1 and every Hilbert space a Banach space of type 2 {{cite:2de2df9dce6f97231faf89c1a591ceee6ef5c3c7}}. Moreover, {{formula:ffd65efb-5e2d-452e-9bf9-50abff28eb86}} spaces are Banach spaces of type {{formula:25265ae6-b07c-4147-9a38-b18186a5ca22}} for {{formula:1d158963-563f-44... | m | 4ba4eb75742109052289d9f66c77fdec |
Understanding what happens to rocky planets and their atmospheres in the habitable zones (HZs) of
low mass stars is currently one of the greatest astronomical challenges. The nearest Earth-mass
planets in the HZ orbit M dwarfs, and these are prime targets for spectroscopic atmospheric
characterization in the next decad... | i | ccb5e9eb46a7d2fb9663aed4c1c700ff |
To numerically reproduce the optical response of our fabricated TiO{{formula:180bb667-37c1-4514-85dd-1bb9798afab6}} nanodisk arrays, we apply the first Born approximation,{{cite:e1feb7b5ebb601d1a6dba53769e8eade24730894}} i.e. the exciting light field for each scatterer is assumed to be the unperturbed incoming light f... | m | 38a55acb194b7d256c68fb791d61d1df |
Defining a realistic augmentation function is an ill-posed problem, and to the best of our knowledge, has not been explored. We propose a heuristic, based on the observation that it might be possible to generate realistic augmented domains that show a small reduction in validation accuracy, when a small augmentation ma... | m | 4af0bbac59a5b483bf419e8febceaba6 |
Similar remarks may be made about the Ni-Na correlation, which has received considerable attention in the literature and is thought to be linked to the production of both elements in Type II SN nucleosynthesis {{cite:96b53e097938a1f0e493c82248cf567130f6e309}}, {{cite:7857791900db92e95f2df62b4605644e84f3e112}}, {{cite:0... | d | 6bef1fe7c664b88c2967ed4b9f9257b3 |
To compare our analytical model of eq. (REF ) against the simulated data we let free nine parameters: four bias parameters {{formula:624dd1f2-27f9-4eb8-8a2e-ab4d90a05c4c}} , three counterterms {{formula:cd399881-72d1-4afb-92bf-72c60a4764a8}} ,
and two stochastic parameters {{formula:4d9d0908-13bd-4262-b30c-2dc7d1f23d02... | r | f6d3d6b07bbb249884ac1740ad835647 |
Amorim and Veloso {{cite:5a567f4df061e392f3bd5ebae834942c6df3b729}} developed 19 features: number of grammatical errors, number of verbs,
number of pronouns, and others. These features fed a linear regression to score an essay.
Fonseca et al. {{cite:d9c581917e6361af92674db3f3434923e544d6b5}} created a pool of 681 featu... | r | 611a1ccd268e2f2280ea55f44f2890b2 |
The basic idea is to control timestepsizes to keep an error estimate at a given tolerance {{formula:f3000ed9-abd8-4e4d-8d6b-65a47515e3ac}} . We use a local error estimate obtained by an embedded technique {{cite:61ffbbbdc6861541681eeb5ab311b276f37c7d4f}}, i.e., for the SDIRK2 method we use the coefficients {{formula:ac... | m | b3ad2f9733392d725ec6e70f664185c0 |
While the Bekenstein bound {{cite:302a5b7841a4e2533714abe11ab3404d72104a6a}}
is fulfilled, our results suggest that the GSL might be violated
not only when {{formula:8527f9bb-891e-4129-ab7c-e1d5a232d33d}} is of the order of {{formula:3a2f7a12-d6f1-4538-946a-6ae5b4ed685b}} but also when it is
larger since the temporal... | d | 6da1413e166519aad8f6605fb2501d78 |
In {{cite:a2756cd2cd9fadcd3acaa1afde9d08cf7455f440}}, Etingof, Schedler and Soloviev introduced the so-called retract relation, an equivalence relation on {{formula:10516eae-cc46-4145-a30c-3ed3c058901c}} which we denote by {{formula:3f388014-0f2d-492f-8fa6-6b5fcbeecec5}} . If {{formula:25b92dc5-5824-453d-b0e4-35e3487e... | r | 3340557277368acdaaf4cf1ff0eaee7a |
Neural machine translation (NMT) systems map text from one language into another via a neural network. Several approaches to NMT have been developed, commonly consisting of an encoder-decoder architecture and an attention mechanism {{cite:c87d6ee893728e7ccf53be7f7bac4407219344e8}}, {{cite:5a59e9902f6cda194a8d7d42bc9a9a... | i | e81e3921dc64f7a736b4ca4b4c9d0d10 |
The complexity of Arrival is particularly interesting in the context of other games on graphs, such as Condon's simple stochastic games, mean-payoff games, and parity games {{cite:6c529c1a1edbeb5dbff40c18cea443d8de8cc9a0}}, {{cite:2b2f306c37a96e48980d396fe24e2db2fd1896ae}}, {{cite:eba29c3d1edaabde41491e3b5fb2d7f4ef1ba0... | i | 13c5d0ea62eb097419e3834ff1743629 |
Existing approaches to meta-embedding rely on self-supervised learning such as autoencoding {{cite:403350c6688230d2e2c791e7db82ec90cacec453}} to find a lower-dimensional hidden representation of the set of source embedding (further discussed in § REF ).
This can be advantageous in cases where (1) pre-training is expens... | i | 2c97a71361b7534dec5325a9742391d3 |
Besides the multiscale challenges, practical applications of the linear transport model usually contain uncertainties {{cite:7a16dab47099c91564cfe8acf7b5c60124e60222}}, {{cite:a2ddf73c0fbb48188912d4523f4945815d58a087}}, {{cite:a05e1b1c5a831d2b0b59c14799a956514dd7509c}}, {{cite:f3991d7c8da70301c7c3ab3e0d00c4b7d2012dcb}}... | i | 785cac1e8326d0dec524e2f29ee4923d |
Furthermore, the semiconductor coupled dots or hetero-dimers has similarity with molecules in molecular electronics{{cite:e334cde3ce5d2f53a9b69341d0e3e32d1a1c03c3}}, {{cite:096e0ad260cf3ec140e1ce4b981bb090e6d7c916}}, where a single molecule with a donor part and an acceptor portion connected by
a bridge acts as a molec... | i | 02653ba351c64e3114a0802cecf269f8 |
We examined the viability of reheating via the evaporation of Q-ball derived black holes in the event that they come to dominate the energy density after inflation and then decay to radiation at later times. We find that this reheats the Universe at a very low temperature {{formula:88bc50ba-319d-4471-8457-fc5b1526e288}... | d | d1c3b720b59d6569bc11b94d813924b8 |
Another issue to be considered
is
that the dependence of the frequency of the gap edge
on the directions of external fields is strong
for the clean case
in the surface resistance. {{cite:a7d072122a19228e02244d8f8677ece2cc37b9c2}}, {{cite:40d97cc1eb6737dfdd6827ad02dded176138fb5e}}
In our calculation,
there is no differe... | d | 57421cb3ae824ea8657ccacabdba0e83 |
Applying Lions' compactness theorem (see Ref. {{cite:703c268422f35821fbb00bdb822320fddf0574b3}}, p. 58) we deduce
that
{{formula:4d48bcf6-b6f9-4568-a337-5e89cf9bd169}}
| r | d55ff0e7e851e5481fa860e0668f2c74 |
As for the computation complexity, we performed FLOPs {{cite:b44aeaf2017b3ad1d9f8d6f54aedb2650d39aeb8}} analysis on the proposed method in four different situations. The results are shown in Table REF . As can be observed, the proposed model with ROI-Seg achieves the lowest computational cost in Frustum US, while it ha... | m | 7ff91aa88a11df4cca3ec4acb9e29bb0 |
In this section, we introduce the concept of a middle Roman dominating function to study the Roman domination number of the middle graph {{formula:f337eb45-5340-48a3-963d-41673063cfd2}} for a given graph {{formula:2b36e3ce-f309-4aba-be94-b3559f74a7f3}} .
A middle Roman dominating function (MRDF) on a graph {{formula:d... | r | b0160e84144e0d3ccfdf489c999a02d1 |
Gravitational torsion may be viewed as a twisting of spacetime, analogous to the curving of spacetime accomplished by curvature tensor. In this regard, Einstein-Cartan-Sciama-Kibble (ECSK) theory of gravity turns out to be convenient {{cite:7308b8a68861f667444fbd58d4d9800c0c2638d5}}, {{cite:3731e25864cda63fe3966fd20b23... | i | 82406cf93a322114f26f5afe4368cfa2 |
In recent years, high-fidelity image synthesis has significantly improved by through the use of Generative Adversarial Networks (GANs) {{cite:93ee6f88019153721a301f79754cc9f22c1d7f1a}}. Whereas early work such as DCGAN {{cite:f1f7429b9083614993c648662edf40492ca3f2bb}} could generate images having a resolution up to 64x... | i | a78c12d27107f7eb4654a9029a5f6cca |
In the subsequent large-scale experiments,
we list the performances of the comparisons in Tables REF and REF . The symbol “–" in Tables REF and REF means that Gurobi fails to solve the problem due to excessive memory requirement. In the implementation of the ADMM, we apply the Sherman-Morrison-Woodbury formula {{cit... | r | c36dce611c5334eed1b5a7f50ba84497 |
where {{formula:c3b26e4a-4c77-4c90-b081-d6eebaca6092}} follows a Gaussian prior, {{formula:3ee5eab9-e2aa-4ec1-9b10-f567c37122a6}} and {{formula:816a0ec8-f8ba-4653-b3cb-5a4fbfcc9147}} are learnable
parameters. With a ConvGNN as the encoder and a simple
multi-layer perception as the decoder, GraphVAE outputs a generat... | m | 7a394155398e8759f334704b5c9c0b28 |
Convolutional Neural Networks (CNNs) have made extraordinary progress in various computer vision tasks, with image classification as a most representative one.
The trained models generally perform well on the testing data which shares similar data distribution to that of the training data. However, in many practical sc... | i | c14a9d00e9fea8e9e1d5b334736ff49c |
As we all knows, the memory effect of the environment is caused by the long time correlation between the system and the environment and the environment can be divided into Markovian and non-Markovian types according to the memory effects. Many methods for quantifying non-Markovian features have been proposed, such as R... | i | 12ff55e01ebb13016f5f8f3ab1b30def |
Baseline Plus. Inspired by several improved baseline methods {{cite:3ce2e2a83787fe9804967d4181ecedb7281c595f}}, {{cite:b104d89eee6fcd1d73ec81db4603c933d105ad9c}}, {{cite:31ea9c2d190f314daa64b71610693623e134c419}}, {{cite:fabc70aafe4ec35a27b56a1a10ef222bbc98b983}}, we propose our modified baseline for few-shot video cla... | m | f46ae3169bef90df8d08190ce019d64f |
We calculated spin susceptibility
and found the rapid linear increase at low temperatures
compared with
that of {{formula:f6de88cf-3153-45d8-9a2c-80947021bc76}} -(ET){{formula:696b87b2-18ec-4aaf-8eb9-3fc8958ffe02}} I{{formula:4ae9691f-a0d7-49fd-9be4-d8b0862130d2}} .{{cite:4eb301f241f47a86cd33f291b00dd12fe24cb767}}
Such... | d | fed48b30382902fe894fb6fe2301917d |
The miniImageNet dataset {{cite:05c362a4e620513612cf2d8bea3bc9a67031f49e}} is a standard benchmark for few-shot learning algorithms for recent works. It consists of 100 classes randomly sampled from the ImageNet; each class contains 600 downsampled images of size 84x84. We follow the widely-used splitting protocol prop... | r | 7b0490b46557c4adf5e95c6657d8c34c |
was obtained in Ref. {{cite:a80b947ec7beeb864d828656b43c0ee1df0d19e6}} taking into account the KSFR relation. With this result Eq. (REF ) has a {{formula:d0f6470c-9906-48d2-bcd9-bab2c608582b}} -resonance pole at {{formula:91a36231-8083-44a4-a03b-0824db81a9d6}} GeV{{formula:fd90c706-3aa8-49ee-8f3b-47677f5749e4}} , whi... | d | a9a747e76571f5d82d516832447c8886 |
The form of {{formula:5849ba17-f282-45be-b757-cf7cc248d69a}} generalizes for other values of {{formula:7c6c6c72-5657-4cbe-ba77-6ccb850180a0}} and {{formula:779d5285-d159-4425-87a4-74503f3d978a}} . The structure of Toeplitz matrix {{formula:39d01352-3c7b-4dc6-933e-6ec70e03f942}} depends on how the vector convolution ... | m | a85c7a1c51d080e05fe4cd3378bbfae3 |
One of the first theories of nonlocal elasticity is the strain-driven model introduced
by {{cite:cdb2a15d6ef68f9963b7a39c9108d87e18abbf15}} in the wake of seminal contributions by {{cite:660f156463f29b8e38ff38c06f9cfe8b3176879b}}, {{cite:e46dcde71bb636bf20d7660aa2b99c5695db769e}}, {{cite:19d071a1b99dae7ec40f3924ea92b03... | i | 6aa9f14eac0c8d25995420215e6ba6a6 |
The classical AIC and MDL methods are based on information theory and these methods have originally been applied to one dimensional data. According to {{cite:c589802aa27f99482c626477cffa9c1dae8d8bfe}} the AIC criterion is given by
{{formula:727a3e8b-0aad-473b-b7dd-fded24616f88}}
| m | f0bf4160a8d6b5a7f1e64b770d42b6b4 |
Table REF contains the obtained results.
Our unaugmented model outperformed other published results, including BPL {{cite:e42ce868fd7c55cc46179e5929f14a2a41162746}} and Prototypical Networks {{cite:2a3fcc1069d0a9693e86d56f3ca089e6cf0a64a2}} and achieved perfect accuracy after applying Hungarian algorithm at test time,... | r | 8e7723a34cf9101f762a7ca3d1460f98 |
Besides, since the combined patch embeddings only contain part of the information in the whole image, pulling the partially combined patches closer to the target view that contains the whole image information is more challenging than pulling the original image pairs and implicitly increasing the asymmetric of the netwo... | d | 80166d8c71e6f15313218cd42f910e1c |
There are a number of extensions of these results in the commutative setting esp. the so called hyper-Kloosterman sums {{cite:13d49e57dd63b0849b3f732538e5fb9c1f751bc6}}, {{cite:b807f065daccbd7474ad8b3115ec41043ad30128}}, {{cite:c77c523674dfde3a5647822f59ccb6f1c6bed097}}, {{cite:1cd83b15f88a3a13d9d17ed1b0366b6544af0aaa}... | i | 71ebd95ffc370ec7043c82a93238ed43 |
Performance of different unlearning methods. Next, we test various unlearning schemes combined with GST. In this set of experiments, we sequentially unlearn one node from each of the selected {{formula:1d14dd03-19c1-4be8-b270-ca9313b923ff}} training graphs. We compare our Algorithm with the unstructured unlearning me... | r | 0e828a98746a66720df869890ba7a42c |
This is the first coreset construction with an optimal dependency on {{formula:921f849c-cea7-4c68-a12a-b4abf1d1a95c}} , at the cost of a quadratic dependency on {{formula:88192f72-f687-42dc-ae6f-a8d362767cd7}} . Previously, all upper bounds either had a dependency of at least {{formula:ad8498cf-f52c-43dd-9190-3be8abde5... | r | 39f1db29ee93957f38ffeb1383abfc02 |
where {{formula:980d1f46-f8e3-41e0-ae78-526d6d011d48}} denotes the subgradient of {{formula:9a6bc3c1-1069-487b-8033-9db8a1b8790d}} at point {{formula:3274432c-a5f4-4f83-a3f9-b440042843a8}} . It is worth pointing out that
{{formula:23521183-cb26-4aeb-95ed-58ecc3c7068f}} can be calculated from Proposition 4 of refere... | m | 47ff6ae7f9559b18eb938e6bd2d38f53 |
We compare IPE-LIIF with five baseline methods: bicubic interpolation, EDSR-baseline{{cite:40bcc072efd76c03e01ce106c085258e1186a388}}, RDN{{cite:074992fddeafbe47a0746bed5df768a09fec8df0}}, MetaSR{{cite:b179cb29d6e3e6e453db99544036d93697b2e443}} and LIIF{{cite:1924daeed416ac3f86ccb7b025d1f3934a43ca59}}. Note that EDSR-b... | r | 6d53505cc0825ff036b79b6584fa866f |
It is known that the irregular set associated to a uniquely ergodic dynamics is empty (cf. {{cite:26da4fb4879741be2ea1d561c602e912f3f992ad}}). Besides, some of the most interesting known examples of minimal non-uniquely ergodic homeomorphisms have zero topological entropy. Thus, in this setting, it is useful to describ... | r | 41309c17b39bc5160a24f3494c9c0101 |
This research makes a first attempt to design the coarse trajectory for the energy minimization in UAV-enabled wireless communications with latency constraints. The proposed approach can be extended to a fine trajectory, e.g., waypoints based on VBS placement and convex optimization (WVC) {{cite:d960ed2e756ec5c0bdaa49d... | d | 0c294c322244a400bbf16ec664f6adae |
is well defined under the energy-momentum conservation principle; see {{cite:3bf6e928f7676a3c578d4948bbb402846be82754}}, Lemma 3.15.3. Let's compute the numerator {{formula:ac386a64-c784-40cf-84c8-9250360bc4b6}} of {{formula:f6494460-1391-4543-b0bb-6caba156f5ce}} .
{{formula:f76c1e5d-390b-4e2f-add4-b8383d5e4e6a}}
| r | bfcb412f02766e9bba44ed318ea883bf |
These results can be found in many convex optimization books and papers, e.g. in {{cite:5a638c2e14a5f45aabe620d6fa90ad1f4056ae6c}}.
| r | 875f302bb21986b0258bd02e2351d220 |
To evaluate the generalized ADMM, we deploy the setup of Matlab codes from {{cite:e6f83ffd87565235f34ec4ceaa76b0b3f15bb385}}, {{cite:6682b70bcc1092bfc165aa9a6876d4b56bca71ab}}. We evaluate the performance in terms of numbers of iterations for convergence and we define {{formula:5a458b63-d0d6-43a0-b302-ee5253413c12}} w... | r | bbb65738c826c2a1d52f99bbc670f327 |
The analysis in {{cite:eff8752ba718f343513da980f91a904cf52728d0}} was dedicated to maps from smooth source spaces with values into locally finite Riemannian simplicial complexes, with striking applications in Geometric Group Theory. Korevaar-Schoen {{cite:73e2444c9dcae3fd5d62bd365d12b4919fca9c31}}, {{cite:afafb8eadf20b... | i | 7634fc8bff5cd7d1d48beb2af27d9009 |
We did not apply any post-processing on the prediction segmentation maps. However, connected component analysis (CCA) or conditional random field (CRF) are among the post-processing approaches that were used in the literature to remove the false positive detections {{cite:240402240012745fa2920ef64c3dda55d075dbab}}.
| d | 3361135c54f118d0c6c4c00b677a5102 |
Condensation of pions, kaons, and diquarks in QCD and QCD-like theories has also been studied using chiral perturbation
theory
({{formula:5f4695b1-46ce-42be-9f2e-571b8a2ef26d}} PT) {{cite:b14f168753101d2df1b86dc8122f4edd71a5d32a}}, {{cite:95c44b56d4e9e4eb098674023d75d4884c8a6d30}}, {{cite:0507c4cea2efe0fdb938545e4f4cf7... | i | 9a6fd8bb8735861a460a1368e339cdf6 |
Finding first integrals is often a complicated task, and during the
past decades several algorithms to construct them have been
developed. One of the most effective methods to produce first integrals
of a given mechanical system is the so-called Lax pair representationOften {{formula:0c0bed1e-2c31-457b-a7e4-c4a82a85456... | i | 6cbd279166a6514e896e0db796ce8cc0 |
The results on the full Nordland dataset show that our proposed CNN+RNN model is capable of learning meaningful temporal relations from a single image sequence of a large driving dataset; while significantly outperforming classical sequence-based methods in runtime, accuracy, and computational requirements. We used a s... | d | 0988dac274f0e902e4a5335065752361 |
Texture is ubiquitous and provides useful cues of material properties of objects and their identity. Visual representations based on orderless aggregations of local features (e.g. SIFT {{cite:25659a8bb30ea83dd77d232cc4ba38e0a0690da1}}, SURF {{cite:dd810ac431304666e40afb0d9ccb1652c0bac41d}}, HOG {{cite:0177cfcfb3a6dc6a8... | i | 40bf3cc005dc5fb4947163f252a5ae47 |
To make NMT models better learn from and cope with lexical constraints, we propose to leverage attention modules {{cite:8fa86d6af9fba1f783d0ee332be4f1e3fddf6789}} to explicitly integrate vectorized lexical constraints. As illustrated in Figure REF , we use vectorized source constraints as additional keys and vectorized... | i | a63c5169a8199f5d1818e2993d4fae33 |
We further report the experimental results on CamVid {{cite:b208b8ea92fcaa4e111c591ca38ec54c8f16081a}} in Table REF , using DDRNet-39 backbone network {{cite:951fd7b524589149bd21dd72c3543ad1d8b5cc85}}. The results show that our method can achieve 70.2 FPS, achieving 18.8% gain over the base model. We increase the FPS w... | r | 9b5733ebaac092b1ca7b9c47e4d46e1d |
Before evaluating CERT, we would like to evaluate our base model PyCodeGPT on HumanEval {{cite:6102d242d74bdf5f16324313920fe7f543f4359a}} compared to several advanced pre-trained models. As shown in Table REF , PyCodeGPT (110M) achieves competitive {{formula:c6e299df-88c6-4305-83fc-1ca9ae4880a5}} pass{{formula:3f16c17... | r | 7093ede9fa759bad9ae3556306e2708b |
it is clear that {{formula:2005e111-8391-4025-be01-49e16d09dd56}} and similarly for system B with equality if and only if all {{formula:f329ed63-2c45-46e1-af78-af8b3181364e}} are equal, where {{formula:6ff9b372-67ff-4cda-bec4-c153e76884e7}} are the eigenvalues of the density matrix {{formula:4a191127-6729-4c4a-8f76-... | d | 4336df19439a0bba4b72374427db80dd |
We compare the performance of two architectures across all datasets. First, we utilize the U-Net architecture {{cite:3074e99fba318c3e7ef857ef8928f90bd2e1899e}}. Our U-Net implementation takes as input the adjoint of the measurements under the forward model {{formula:48209ce1-cdb4-43f2-bcc1-293ae5436d64}} , which is the... | m | ae239a4bb60f073948e50b8469b678cf |
BiLSTM-CNN-CRF
We used the PyTorch implementation of a BiLSTM-CNN-CRF model {{cite:775bdf6f8bddd1bc8bf7fd22ce9f5306c27ccc08}}.
The model combines the word embeddings with the character-level representations extracted using the CNN and feeds them into the BiLSTM module with the CRF output layer.
Word embeddings are usua... | m | bc90b155c55ff6648eafcf7929915abd |
We do not consider the gravitational lensing by a marginally unstable photon sphere for {{formula:7db812e7-d73a-4b87-900d-64f6e19c7441}} .
In this case, we cannot use Eqs. (REF ) and (REF ) since the deflection angles do not logarithmically divergence in the strong deflection limit.
We can use formulas in Ref. {{cite:b... | d | fb5a7fd4d407b29e61c43df7caf7b526 |
We named our AI improviser A.L.Ex, the Artificial Language Experiment, an homage to Alex the Parrot, trained to communicate using a vocabulary of 150 words {{cite:8fab90e1281f2d1dd3068ecff32562360188edda}}.
The core of A.L.Ex consists of a text-based chatbot implemented as a word-level sequence-to-sequence recurrent ne... | m | 4b0d163ae08165457fb6400e886b93be |
Table REF and REF show examples of incorrect predictions when a counterfeit key is embedded into the recurrent neural network (RNN) model during inference phase. For classification tasks (i.e. TREC-6 {{cite:afb789cb0662b3b6c08d05fb38af692e60609c77}}), Table REF shows that when a counterfeit key is used, the RNN mode... | r | 5f1258e8bef0154b2bc859292cba767a |
The emergent rule for graph growth is related to fractal properties of the graph. Structures present in the seed or formed early on in this process will determine the basic fractal structure. A growth algorithm similar to the iterated tensor products might satisfy this property.
The emergent rule for graph growth is ... | m | ce91aff71cd2677672c7ef07b63b62d0 |
Compared to the rich literature on the well-posedness of classical water waves, the research on the well-posedness of water waves problem with non-smooth boundaries (we call it “non-smooth water waves”) just started several years ago and there are a lot of open questions . In general, there are two kinds of non-smooth ... | i | 1daf66fbfb117f9c6fc5c017fe79c499 |
The following definitions of a blr makes use of the introductions in {{cite:cc7c91e0f71c793470233738c00036244cd02d04}}.
In contrast to a deterministic perspective to learn the model weights of a linear regression model, a Bayesian approach gives additional insights through the posterior, especially when there is insuff... | m | 56a5050a12b1c27d1d188b726624b598 |
Fig. REF shows the minimum on-line training overhead versus {{formula:927b490c-0aa4-4049-b49f-b4e0de0cca6f}} , with {{formula:b507352e-940a-4537-b7be-94cf979360b7}} and {{formula:be1c48b0-1dec-4bb6-9d19-754fc14c4734}} . On one hand, it is observed that both our proposed schemes are more efficient than the scheme in {... | r | 78e139cc69f66c68152716c6af8c98b2 |
A widely used technique to study the stability of the synchronous solution is based on the derivation of a Master Stability Function (MSF) {{cite:438ddcff193685b932990df5a7d9f87a2cc18d04}}. The technique consists in linearizing Eqs. (REF ) around {{formula:a6d49a2a-99e0-401c-bbaa-2565160cacb1}} , and then applying a pr... | m | f7b21d9571c5edf64f6806c458474d85 |
Split MNIST and Permuted MNIST: Subsequent classes in the MNIST {{cite:2f348876d5c71ce9c6052da1104b7a510256efcc}} dataset are paired together and presented as a task in Split MNIST, a well-known benchmark for continual learning. This results in 5 incremental tasks. In Permuted MNIST, each task is a unique spatial permu... | r | 06277e88dbc963558ab80a64d867e4fa |
In section REF , we outlined the most conceptually similar methods that conducted large-scale model pretraining with task-agnostic parallel sentence alignment as part of the training routine {{cite:6728b9c079e5b3ab9140bd3d1f0593928e7deab0}}, {{cite:0032534a03dded45256c6b11289ae71f16efd301}}, {{cite:7f00783d26e63b4c1415... | d | c0658b195b9af6f20c18b68f07f05130 |
End-to-end text spotting results on the Total-Text dataset are reported in Table REF .
Compared with the previous fastest ABCNet {{cite:40c4b988bf4ea4e44b12ef709293e29fd25513ff}}, our method with shorter side being 640 pixels runs 1.3 times faster, while our end-to-end text spotting F-measure is 2.2 points higher (66.4... | r | a040f5f56bea4295d25e0d7aedee5158 |
In {{cite:7ec4a8176c66a7f4dffb42f0309784a7a0d3d8cb}}, the optical and thermal information was mined to recognize pedestrians. In {{cite:f4b5bdd10581fcbe48758c31106e76537be07bcc}}, a cross-view cross-scene multi-view counting method was proposed to improve the performance and generalization of the model. In {{cite:e0fd5... | m | 72f1a0d4a831ec9020a49a4ba3ac4d78 |
where {{formula:80034e81-7725-4cb2-8fc0-c9026260245c}} denotes a {{formula:ec160c6e-c876-47db-9495-b7c355f8c049}} -dimensional random vector taking values as in the rows of {{formula:0286193e-9493-47c1-957e-ac03bedf4c45}} . In this case the random compositions {{formula:ae7ee203-a4f2-4ad0-bcaa-9b85cb75d8e2}} are dist... | i | d0a47e7f63cfdf719b368ee94c34b009 |
Fredholm integral operators canonically arise in numerous applications ranging from a classical fixed-point formulation of linear elliptic boundary value problems {{cite:3f8d8663b34557620df6068a18d3c166168730a3}} to Fréchet derivatives of recent models in theoretical ecology {{cite:063e4dfd2b4ec2e8053bb745784025420e21b... | i | 7227a8abd8691c95dddf13235db49965 |
Based on the nature of problems, a ConvAI system is expected to solve three major research problems {{cite:0fea7d98567a468df900c615081f63b904fd14ca}}. Question Answering (QA) involves providing answers to user queries through conversation, using the knowledge drawn from various data sources like a snippet from a text, ... | i | 4718c1f0a1b41a2c226b0add2c661fd6 |
blackSo far, much attention has been devoted to the detection of binary mesoscale structures, i.e. communities and, to a far less extent, core-periphery structures: the efforts to solve these problems have led to a number of approaches that are briefly sketched below (see {{cite:a9caa62859934c2454ef4daa9f76a20962fe4dd1... | i | ff283010e96c72186df2945d966aa2e6 |
We have shown that {{formula:d7f6e74a-e15f-4369-99f3-a47232ad7682}} scalar field perturbations of the nearly extremal black holes are governed by the effective potential which has a deep negative gap, and, that, nevertheless, time-domain profiles are decaying, which points to the stability of the scalar field.
When ... | d | aee24e1a33db91fce12e8290746573a9 |
LASER+LSTM+T5: this method uses pre-trained LASER embeddings, which are passed as input to a LSTM model for category classification and then to a T5 model for summarization;
C-SKIP {{cite:5e96aac49213844236116bd86d679a5ab5357212}}: this is a centroid-based method using a FastText skipgram model trained on the CrisisL... | m | c2724127b4b3ca817b73e52e8db91d13 |
To show how general the proposed framework is, we propose to express the state-of-the-art approach proposed by Ustun et al. {{cite:e232155acd1a99c9fc868e828df738f3852a1387}} in the proposed formalism
highlighting the definition of the three involved functions.
In this approach, the user knowledge {{formula:8a8cc88d-dad... | m | 9dbb330e676348354702d1982653bfc8 |
DL has become very popular for imaging inverse problems {{cite:30f3fca7715b8d6b38c667194addea39e5db22ab}}, {{cite:97493874f407c4340994d8ba5147c5924737d72a}}, {{cite:d20b61ccd53ceae9b08c91032684dd41933cea5c}}, {{cite:afeeabd6938edb435f83fe2540aa6131cb9d9795}} due to its excellent performance.
Traditional DL methods firs... | m | 5f9713ca7ab9d20a53c3f3c7dc4ffbac |
Despite the growing application of GNNs to fMRI data, a major limitation is the requirement that the graph adjacency matrix is known prior to model training. It is therefore implicitly assumed that FCMs represent the ground-truth dependency structure of fMRI data. In reality, however, this is not the case, as FCMs are ... | i | 9602aeb681da11938d4faa59cd41c925 |
We note that {{formula:1a96fbce-dcba-4913-a44b-45c3735c9278}} , {{formula:d38258bf-77b0-45c8-aaab-d301c30cfa80}} by the regularity assumptions of Definition REF . (See {{cite:367d00021325a5f0061db69111dafbbf59552d5d}}, for example.) Thus (REF ) is well-defined.
| r | b74d19539e9147c5abcf6fd0e193a753 |
In order to investigate the learned feature distributions quantitatively,
Figure REF provides Proxy {{formula:79a9ecce-382a-4c44-b592-ed7af8297edc}} -Distance (PAD) between the feature distributions from STA, ROS, and UADAL.
PAD is an empirical measure of distance between domain distributions {{cite:8867d57187a9b22273... | r | d1efe1a3b2727be550ddbaed1c1cffe9 |
To train MIME, we built a new dataset called 3D-FRONT Human that extends the large-scale synthetic scene dataset 3D-FRONT {{cite:983504f9c1a7b93d62f90c0aa2049d1c317438dc}}.
Specifically, we automatically populate the 3D scenes with humans, i.e., non-contact humans (a sequence of walking motion and standing humans) as w... | i | 450637a51d858d8db485aca8825aaaa0 |
where {{formula:a014fa9a-7f78-4a10-9bf1-f33296a15cbc}} and {{formula:f362cda2-b762-4344-a738-d4df405aef33}} refer to
the average path-length of high-{{formula:a4b58798-96ab-4293-9520-71127163a1b9}} particles in the in-plane
and out-of-plane directions. For every temperature profile, {{formula:763a2e8b-f93e-4428-b77... | r | 0219dc0307470a6e49f171dd14a31f4e |
Body2Hands w/wo image{{cite:4bd7a2ae8f416101c56dad01b5ec3e13cfc2ddea}}: State-of-the-art 3D hand shape synthesis and estimation by a learned deep prior of body motion. Models are re-trained using the codes released by the authors, since the pre-trained model is not available.
Body2Hands*{{cite:4bd7a2ae8f416101c56dad0... | m | c1291d2dfde26c6aef7f867769465482 |
In this work we analyse the combination of differentially private training, model compression and adversarial training techniques against model poisoning attacks. We determined that for the strongest insider adversary, post-training quantization did not have a significant impact on the results of the attack. The opposi... | d | 16c7c60c53e94288e8d1991a386808f3 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.