The compliance of soft robotic arms renders the development of accurate kinematic & dynamical models especially challenging. The most widely used model in soft robotic kinematics assumes Piecewise Constant Curvature (PCC). However, PCC fails to effec
tively handle external forces, or even the influence of gravity, since the robot does not deform with a constant curvature under these conditions. In this paper, we establish three-dimensional (3D) modeling of a multi-segment soft robotic arm under the less restrictive assumption that each segment of the arm is deformed on a plane without twisting. We devise a kinematic and dynamical model for the soft arm by deriving equivalence to a serial universal joint robot. Numerous experiments on the real robot platform along with simulations attest to the modeling accuracy of our approach in 3D motion with load. The maximum position/rotation error of the proposed model is verified 6.7x/4.6x lower than the PCC model considering gravity and external forces.
Random uniform sampling has been studied in various statistical tasks but few of them have covered the Q-error metric for cardinality estimation (CE). In this paper, we analyze the confidence intervals of random uniform sampling with and without repl
acement for single-table CE. Results indicate that the upper Q-error bound depends on the sample size and true cardinality. Our bound gives a rule-of-thumb for how large a sample should be kept for single-table CE.
The performance of fine-tuning pre-trained language models largely depends on the hyperparameter configuration. In this paper, we investigate the performance of modern hyperparameter optimization methods (HPO) on fine-tuning pre-trained language mode
ls. First, we study and report three HPO algorithms performances on fine-tuning two state-of-the-art language models on the GLUE dataset. We find that using the same time budget, HPO often fails to outperform grid search due to two reasons: insufficient time budget and overfitting. We propose two general strategies and an experimental procedure to systematically troubleshoot HPOs failure cases. By applying the procedure, we observe that HPO can succeed with more appropriate settings in the search space and time budget; however, in certain cases overfitting remains. Finally, we make suggestions for future work. Our implementation can be found in https://github.com/microsoft/FLAML/tree/main/flaml/nlp/.
We propose the ChaCha (Champion-Challengers) algorithm for making an online choice of hyperparameters in online learning settings. ChaCha handles the process of determining a champion and scheduling a set of `live challengers over time based on sampl
e complexity bounds. It is guaranteed to have sublinear regret after the optimal configuration is added into consideration by an application-dependent oracle based on the champions. Empirically, we show that ChaCha provides good performance across a wide array of datasets when optimizing over featurization and hyperparameter decisions.
This paper proposes a novel deep learning-based video object matting method that can achieve temporally coherent matting results. Its key component is an attention-based temporal aggregation module that maximizes image matting networks strength for v
ideo matting networks. This module computes temporal correlations for pixels adjacent to each other along the time axis in feature space, which is robust against motion noises. We also design a novel loss term to train the attention weights, which drastically boosts the video matting performance. Besides, we show how to effectively solve the trimap generation problem by fine-tuning a state-of-the-art video object segmentation network with a sparse set of user-annotated keyframes. To facilitate video matting and trimap generation networks training, we construct a large-scale video matting dataset with 80 training and 28 validation foreground video clips with ground-truth alpha mattes. Experimental results show that our method can generate high-quality alpha mattes for various videos featuring appearance change, occlusion, and fast motion. Our code and dataset can be found at: https://github.com/yunkezhang/TCVOM
We identify 225 filaments from an H$_2$ column density map constructed using simultaneous $^{12}$CO, $^{13}$CO, and C$^{18}$O (J=1-0) observations carried out as a part of the MWISP project. We select 46 long filaments with lengths above 1.2 pc to an
alyze the filament column density profiles. We divide the selected filaments into 397 segments and calculate the column density profiles for each segment. The symmetries of the profiles are investigated. The proportion of intrinsically asymmetrical segments is 65.3$%$, and that of intrinsically symmetrical ones is 21.4$%$. The typical full width at half maximum (FWHM) of the intrinsically symmetrical filament segments is $sim$ 0.67 pc with the Plummer-like fitting, and $sim$ 0.50 pc with the Gaussian fitting, respectively. The median FWHM widths derived from the second-moment method for intrinsically symmetrical and asymmetrical profiles are $sim$ 0.44 and 0.46 pc, respectively. Close association exists between the filamentary structures and the YSOs in the region.
We have studied the properties of molecular clouds in the second quadrant of the Milky Way Mid-plane from l$=$104$.!!^{circ}$75 to l$=$119$.!!^{circ}$75 and b$=-$5$.!!^{circ}$25 to b$=$5$.!!^{circ}$25 using the $^{12}$CO, $^{13}$CO, and C$^{18}$O $J=
1-0$ emission line data from the Milky Way Imaging Scroll Painting project (MWISP). We have identified 857 and 300 clouds in the $^{12}$CO and $^{13}$CO spectral cubes, respectively, using the DENDROGRAM + SCIMES algorithms. The distances of the molecular clouds are estimated and the physical properties like masses, sizes, and surface densities of the clouds are tabulated. The molecular clouds in the Perseus arm are about 30$-$50 times more massive and 4$-$6 times larger than the clouds in the Local arm. This result, however, is likely biased by distance selection effects. The surface densities of the clouds are enhanced in the Perseus arm with an average value of $sim$100 M$_{odot}$ pc$^{-2}$. We selected the 40 most extended ($>$0.35 arcdeg$^2$) molecular clouds from the $^{12}$CO catalog to build the H$_2$ column density probability distribution function (N-PDF). About 78% of the N-PDFs of the selected molecular clouds are well fitted with log-normal functions with only small deviations at high-densities which correspond to star-forming regions with scales of $sim$1-5 pc in the Local arm and $sim$5-10 pc in the Perseus arm. About 18% of the selected molecular clouds have power-law N-PDFs at high-densities. In these molecular clouds, the majority of the regions fitted with the power-law correspond to molecular clumps of sizes of $sim$1 pc or filaments of widths of $sim$1 pc.
This paper proposes a novel active boundary loss for semantic segmentation. It can progressively encourage the alignment between predicted boundaries and ground-truth boundaries during end-to-end training, which is not explicitly enforced in commonly
used cross-entropy loss. Based on the predicted boundaries detected from the segmentation results using current network parameters, we formulate the boundary alignment problem as a differentiable direction vector prediction problem to guide the movement of predicted boundaries in each iteration. Our loss is model-agnostic and can be plugged into the training of segmentation networks to improve the boundary details. Experimental results show that training with the active boundary loss can effectively improve the boundary F-score and mean Intersection-over-Union on challenging image and video object segmentation datasets.
We study the hardness of Approximate Query Processing (AQP) of various types of queries involving joins over multiple tables of possibly different sizes. In the case where the query result is a single value (e.g., COUNT, SUM, and COUNT(DISTINCT)), we
prove worst-case information-theoretic lower bounds for AQP problems that are given parameters $epsilon$ and $delta$, and return estimated results within a factor of 1+$epsilon$ of the true results with error probability at most $delta$. In particular, the lower bounds for cardinality estimation over joins under various settings are contained in our results. Informally, our results show that for various database queries with joins, unless restricted to the set of queries whose results are always guaranteed to be above a very large threshold, the amount of information an AQP algorithm needs for returning an accurate approximation is at least linear in the number of rows in the largest table. Similar lower bounds even hold for some special cases where additional information such as top-K heavy hitters and all frequency vectors are available. In the case of GROUP-BY where the query result is not a single number, we study the lower bound for the amount of information used by any approximation algorithm that does not report any non-existing group and does not miss groups of large total size. Our work extends the work of Alon, Gibbons, Matias, and Szegedy [AGMS99].We compare our lower bounds with the amount of information required by Bernoulli sampling to give an accurate approximation. For COUNT queries with joins over multiple tables of the same size, the upper bound matches the lower bound, unless the problem setting is restricted to the set of queries whose results are always guaranteed to be above a very large threshold.
Using the PMO-13.7 m millimeter telescope at Delingha in China, we have conducted a large-scale simultaneous survey of $^{12}$CO, $^{13}$CO, and C$^{18}$O $J=1-0$ emission toward the sky region centered at $l$=$209.7^circ$, $b$=$-$2.25$^circ$ with a
coverage of $4.0^circ times 4.5^circ$. The majority of the emission in the region comes from the clouds with velocities lying in the range from $-$3 km s$^{-1}$ to 55 km s$^{-1}$, at kinematic distances from 0.5 kpc to 7.0 kpc. The molecular clouds in the region are concentrated into three velocity ranges. The molecular clouds associated with the ten H II regions/candidates are identified and their physical properties are presented. Massive stars are found within Sh2-280, Sh2-282, Sh2-283, and BFS54, and we suggest them to be the candidate excitation sources of the H II regions. The distributions of excitation temperature and line width with the projected distance from the center of H II region/candidate suggest that the majority of the ten H II regions/candidates and their associated molecular gas are three-dimensional structures, rather than two-dimensional structures.
