image
stringlengths
42
218
text
stringlengths
100
1k
paper_id
stringlengths
12
12
figure_idx
int64
1
312
https://arxiv.org/html/2…2.04230v1/x2.png
Figure 2:System Diagram forXAttnMark.XAttnMarkconsists of a watermark generator and a watermark detector, with a shared embedding table that facilitates message decoding through a cross-attention module. In the generator part, we first employ an encoder network to encode the audio latent and then apply a temporal modul...
2502.04230v1
2
https://arxiv.org/html/2…gures/mushra.png
Figure 5:MUSHRA subjective listening test results comparing perceptual quality across different watermarking methods. Higher scores indicate better audio quality as rated by human listeners. Our method achieves quality scores comparable to AudioSeal.
2502.04230v1
6
https://arxiv.org/html/2…2.04230v1/x5.png
Figure 6:Validation accuracy and quality curve of different methods over training steps. Using a blended architecture,XAttnMarkis able to achieve a better balance in terms of learning efficiency in detection, and message-bit decoding and watermark imperceptibility. Compared to our method, WavMark, the fully-shared arch...
2502.04230v1
7
https://arxiv.org/html/2…2.04230v1/x6.png
Figure 7:Detection and Attribution Accuracy of our method on augmented samples across different augmentation strengths.
2502.04230v1
8
https://arxiv.org/html/2…2.04230v1/x7.png
Figure 8:The detection and attribution accuracy ofXAttnMarkacross different audio duration and watermark strengthα𝛼\alphaitalic_αon MusicCaps. The performance is averaged over all the standard audio editing. The attribution pool size is set to{100,1000,10000}100100010000\{100,1000,10000\}{ 100 , 1000 , 10000 }.
2502.04230v1
9
https://arxiv.org/html/2…2.04230v1/x8.png
Figure 9:Visualization of the original audio and watermark residuals across four different methods for a randomly selected sample from MusicCaps dataset, showing waveform (left), spectrogram (middle), and mel-spectrogram (right) representations.
2502.04230v1
10
https://arxiv.org/html/2…2.04230v1/x9.png
Figure 10:Visualization of the per-tile masking penalty distribution. The blue regions indicate tiles assigned a lower penalty for embedding a watermark. Our weighting scheme provides more fine-grained control. The audio sample is randomly selected from MusicCap(Agostinelli et al.,2023), and the watermarked audio is ge...
2502.04230v1
11
https://arxiv.org/html/2…2.04229v1/x1.png
Figure 1.Illustration of the threat models. (a) Previous works assume that the data owner may be malicious and inject backdoors into the distilled dataset before distributing it to users. (b) In contrast, our threat model is more practical. We assume the data owner is benign. However, third parties, such as hackers or ...
2502.04229v1
1
https://arxiv.org/html/2…7/imgs/ttsne.png
Figure 3.t-SNE visualization of the feature space. “Stars” and “Circles” represent the concept archetypes and real images, respectively. The reconstructed archetypes align closely with the deep feature representations of real images, effectively bridging the gap between the distilled data and real images.
2502.04229v1
3
https://arxiv.org/html/2…2.04229v1/x9.png
Figure 5.Visualization of benign and malicious distilled data. Without a direct comparison, users may struggle to sense the subtle differences due to the inherent abstraction of distilled datasets.
2502.04229v1
11
https://arxiv.org/html/2…9/images/scp.png
Figure 1:A overall pipeline for SCP. During training, two augmented viewsTasuperscript𝑇𝑎T^{a}italic_T start_POSTSUPERSCRIPT italic_a end_POSTSUPERSCRIPTandTbsuperscript𝑇𝑏T^{b}italic_T start_POSTSUPERSCRIPT italic_b end_POSTSUPERSCRIPTof an image are generated from the dataset and processed by a frozen feature extra...
2502.04226v1
1
https://arxiv.org/html/2…ges/SCP-main.png
Figure 3:The visualization of clustering performance for SCP-CLIP withViT-B/32 backbone.(Left): An example of an image-to-image search on STL-10, showing clusters produced by CLIP (Top) and SCP (Bottom);(Right): Visualization of clustering performance. SCP-CLIP effectively enhances CLIP’s clustering performance through...
2502.04226v1
7
https://arxiv.org/html/2…9/images/pA3.png
Figure 5:Comparison of different loss weightsα𝛼\alphaitalic_α, the solid line is the mean ARI across five runs, and the shaded region plots the standard deviation across our runs.
2502.04226v1
13
https://arxiv.org/html/2…e-search-scp.png
Figure 7:SCP-CLIP Image-to-Image Search Pipeline depicted in Fig.3, utilizing the normalized outputs of the trained cluster headg(.)g(.)italic_g ( . )without applying softmax.
2502.04226v1
15
https://arxiv.org/html/2…/image_cloud.png
Figure 8:The t-SNE representations learned by SCP-DINO on STL-10.A⁢C⁢C:98.4%:𝐴𝐶𝐶percent98.4ACC:98.4\%italic_A italic_C italic_C : 98.4 %
2502.04226v1
16
https://arxiv.org/html/2…e_Cloud_CLIP.png
Figure 9:The t-SNE representations learned by SCP-CLIP on STL-10.A⁢C⁢C:97.9%:𝐴𝐶𝐶percent97.9ACC:97.9\%italic_A italic_C italic_C : 97.9 %
2502.04226v1
17
https://arxiv.org/html/2…0/images/imr.png
Figure 2:Meta architecture for ÉCLAIR showcasing the usage with two different (out of eight valid) prompts: Example a) uses the maximal information prompt to return bounding boxes along with their semantic class, markdown text, and tables and formulas. In b) we ask the model to return only markdown text without boxes o...
2502.04223v1
5
https://arxiv.org/html/2…ng_order_img.png
Figure S1:Illustrations of reading order over relevant text-like elements, i.e. Text, Section-header, List-item, Title and Formula. Other semantic classes (such as Picture, Footnote and Page-footer in the examples here) are not included in the reading order of the main body. (Note: We are not showing all the classes)
2502.04223v1
7
https://arxiv.org/html/2…n_matrix_avg.png
Figure S3:Confusion matrix for ÉCLAIR boxes matched with ground truth on the DocLayNet evaluation dataset, averaged over thresholds ofI⁢o⁢U≥{0.5,0.55,…,0.9,0.95}𝐼𝑜𝑈0.50.55…0.90.95IoU\geq\{0.5,0.55,...,0.9,0.95\}italic_I italic_o italic_U ≥ { 0.5 , 0.55 , … , 0.9 , 0.95 }.
2502.04223v1
11
https://arxiv.org/html/2…llcat_iou0.5.png
Figure S4:PR-Curves for individual classes and averaged over all classes forI⁢o⁢U≥0.5𝐼𝑜𝑈0.5IoU\geq 0.5italic_I italic_o italic_U ≥ 0.5and 1001 recall-bins evaluated on the DocLayNet evaluation dataset. For ÉCLAIR, the scores are taken from the class-token logits. The mean precision and recall are taken fromTab.S1.
2502.04223v1
12
https://arxiv.org/html/2…0/images/ex2.png
Figure S5:Examples of pages with tables, formulae and pictures. On the left, predicted bounding boxes superimposed on the original sample image. On the right, the corresponding full predictions.
2502.04223v1
13
https://arxiv.org/html/2…ty_hierarchy.png
Figure 1.An example of a predefined hierarchy used to raise the abstraction level of low-level event logs, where ”root” is the most abstract event andAssemblyLoader/Start_System_Threadingis the most specific, detailed event.
2502.04219v1
1
https://arxiv.org/html/2…2.04207v1/x1.png
Figure 1:The upper panel shows the general image stitching result and it can be successfully stitched. The lower panel shows the failed endoscopic image stitching result by using the traditional feature-based image stitching method.
2502.04207v1
1
https://arxiv.org/html/2…2.04207v1/x2.png
Figure 2:This figure illustrates an enhanced endoscopic image preprocessing pipeline designed for stitching endoscopic images. The process begins with the input endoscopic images, which undergo ellipse fitting to identify and mark regions of interest. Following ellipse fitting, the images are subjected to rotation adju...
2502.04207v1
2
https://arxiv.org/html/2…2.04207v1/x3.png
Figure 3:This figure presents a boxplot illustrating the number of valid matched feature points for keyframe selection across 20 videos using three different methods. The Original unwrapped images show the lowest number of valid matched feature points. In contrast, applying the AHE method to the unwrapped images marked...
2502.04207v1
3
https://arxiv.org/html/2…2.04207v1/x4.png
Figure 4:Result comparison among different image processing stages. The original unwrapped image shows only 5 valid matched feature points, resulting in a failed stitching outcome. The AHE unwrapped image, with 52 valid matched feature points, demonstrates improved stitching but still exhibits visible gaps between the ...
2502.04207v1
4
https://arxiv.org/html/2…2.04206v1/x1.png
Figure 1:Packet delay distribution for LTT and QLTT in a resource allocation problem (adapted from[2]by using a different calibration data set𝒵𝒵\mathcal{Z}caligraphic_Z).
2502.04206v1
1
https://arxiv.org/html/2…2.04204v1/x1.png
Figure 1:Scatter plots of ASR to the ratio of the square root of the adversarial token suffix length in jailbreak attacks to the adversarial token suffix length during AT (i.e.,Mtest/Mtrainsubscript𝑀testsubscript𝑀train\sqrt{M_{\text{test}}}/M_{\text{train}}square-root start_ARG italic_M start_POSTSUBSCRIPT test end_P...
2502.04204v1
1
https://arxiv.org/html/2…2.04204v1/x6.png
Figure 2:Curves of the ASR versus the adversarial suffix token length during AT (i.e.,Mtrainsubscript𝑀trainM_{\text{train}}italic_M start_POSTSUBSCRIPT train end_POSTSUBSCRIPT) under jailbreak attacks with different adversarial suffix token lengths (i.e.,Mtestsubscript𝑀testM_{\text{test}}italic_M start_POSTSUBSCRIPT ...
2502.04204v1
7
https://arxiv.org/html/2…2.04204v1/x7.png
Figure 3:Utility analysis based on LC-WinRate. A high LC-WinRate indicates strong model utility. An LC-WinRate of50%percent5050\%50 %means that the evaluated model has the same quality as the reference model Davinci003.
2502.04204v1
8
https://arxiv.org/html/2…2.04199v1/x1.png
Figure 1:This figure shows that we expanded our training set from 435 to 7050 images for training EoE deep learning models. The left panel shows that the traditional AI model only uses rare pediatric EoE data. The right panel shows that we employ a wider range of data from different sources.
2502.04199v1
1
https://arxiv.org/html/2…2.04199v1/x2.png
Figure 2:This figure provides an overview of the learning framework. (a) illustrates the image classes, which include six EoE classes and five non-EoE classes, along with their approximate locations in the upper GI tract. (b) depicts the proportion of each class within the entire upper GI dataset. (c) displays the dist...
2502.04199v1
2
https://arxiv.org/html/2…2.04199v1/x3.png
Figure 3:This figure shows the endoscopy EoE image and corresponding attention map. Image a shows an endoscopic image with label “exudate”, whose attention map is presented in image b. Image c presents an endoscopic image with “rings”, whose attention map visualization
2502.04199v1
3
https://arxiv.org/html/2…2.04194v1/x1.png
Figure 1:An overview of GRAPE. GRAPE takes multiple off-the-shelf existing datasets and optionally generates new responses, finds overlapping instructions with multiple different responses and selects responses that align with the base model’s distribution. This dataset is then used for standard supervised fine-tuning.
2502.04194v1
1
https://arxiv.org/html/2…2.04194v1/x2.png
Figure 2:The training loss curve of Llama3.1-8B onTulu-3-Olmo-2combo.Beststands for training with highest-probability responses (GRAPE-selected);Randomstands for randomly chosen responses for each instruction in the pool,Worststands for lowest probability responses in the base model’s distribution. Throughout training,...
2502.04194v1
2
https://arxiv.org/html/2…2.04194v1/x3.png
Figure 3:The average performance curve of Llama3.1-8B onTulu-3-Olmo-2combo.Beststands for training with highest-likelihood responses (GRAPE-selected);Randomstands for randomly chosen responses for each instruction in the pool. The average is taken over all items on table2except AlpacaEval2.
2502.04194v1
3
https://arxiv.org/html/2…2.04192v1/x1.png
Figure 1:The two major research questions we explore: (i) the grounding ability of pixel-level MLLMs in challenging scenarios (first row), (ii) the ability of vanilla MLLMs to perform grounding and when does it emerge (second row). Second row shows the noun phrase and its corresponding segmentation mask, highlighted in...
2502.04192v1
1
https://arxiv.org/html/2…2.04192v1/x2.png
Figure 2:First shortcoming of pixel-level MLLMs is the degraded performance in visual question answering. The Predicted segmentation masks corresponding to the [SEG] token/s are highlighted in red.
2502.04192v1
2
https://arxiv.org/html/2…2.04192v1/x3.png
Figure 3:Second shortcoming of pixel-level MLLMs is the degraded performance in pixel-level visual grounding in certain models. The predicted segmentation is highlighted in red.
2502.04192v1
3
https://arxiv.org/html/2…2.04192v1/x4.png
Figure 4:Third shortcoming of pixel-level MLLMs is the degraded performance in instruction following, where the question is instructing the model to generate one letter from the options.
2502.04192v1
4
https://arxiv.org/html/2…2.04192v1/x5.png
Figure 9:The automatic selection baseline, PixFoundation, which uses a simple mechanism of highlighting the predicted masks in red then prompting a multi-modal large language model to select the right mask from the group of highlighted images, followed by the final mask selection.
2502.04192v1
25
https://arxiv.org/html/2…/292_015_att.jpg
Figure 11:Normalized attention maps visualization showing the noun phrase and its corresponding attention in the output text for two PixMMVP examples using Cambrian-1 (8B) base MLLM. While the attention maps can not be directly used as segmentation, yet it provides initial locations for the maximally attended pixels co...
2502.04192v1
39
https://arxiv.org/html/2…ys/1_000_000.jpg
Figure 12:First example of when grounding emerges, corresponding to Image 1 in Fig.17. Each row has the corresponding noun phrase on top and three potential SAM predicted masks highlighted in red using the maximum attention point of this noun phrase as a point prompt, highlighted as a black dot. It shows the output fro...
2502.04192v1
46
https://arxiv.org/html/2…ys/6_000_000.jpg
Figure 13:Third example of when grounding emerges, corresponding to Image 6 in Fig.17. Each row has the corresponding noun phrase on top and three potential SAM predicted masks highlighted in red using the maximum attention point of this noun phrase as a point prompt, highlighted as a black dot. It shows the output fro...
2502.04192v1
56
https://arxiv.org/html/2…ys/3_000_000.jpg
Figure 14:Second example of when grounding emerges, corresponding to Image 3 in Fig.17. Each row has the corresponding noun phrase on top and three potential SAM predicted masks highlighted in red using the maximum attention point of this noun phrase as a point prompt, highlighted as a black dot. It shows the output fr...
2502.04192v1
69
https://arxiv.org/html/2…/161_000_000.jpg
Figure 15:Fourth example of when grounding emerges, corresponding to Image 161 in Fig.17. Each row has the corresponding noun phrase on top and three potential SAM predicted masks highlighted in red using the maximum attention point of this noun phrase as a point prompt, highlighted as a black dot. It shows the output ...
2502.04192v1
85
https://arxiv.org/html/2…/292_000_000.jpg
Figure 16:Fifth example of when grounding emerges, corresponding to Image 292 in Fig.17. Although the ground-truth segmentation is all background, since theoracleselection sees all the candidate masks with IoU zero, it defaults to the last selection. Each row has the corresponding noun phrase on top and three potential...
2502.04192v1
98
https://arxiv.org/html/2…raw_images/1.jpg
Figure 17:Examples of the noun phrases and concept categories where the grounding emerged corresponding to the provided referring expression as input to the MLLM following the third evaluation protocol. The base MLLM used in these experiments are Llava 1.5 7B.
2502.04192v1
114
https://arxiv.org/html/2…2.04176v1/x1.png
Figure 1.Illustration of the MRAMG task (above), with scenarios below showing how integrating text and images enhances clarity and understanding.
2502.04176v1
1
https://arxiv.org/html/2…2.04176v1/x2.png
Figure 2.The construction pipeline of MRAMG-Bench. It is divided into three stages: (1) Data Selection and Preprocessing, we collect, clean and preprocess data to ensure the quality of initial datasets. (2) QA Generation and Refinement, we formulate questions and answers and refine the formulated QA pairs. (3) Data Qua...
2502.04176v1
2
https://arxiv.org/html/2…s/ls_ag_news.png
Figure 2:Accuracy scores for LS task performance. As token replacements increase, task performance decreases, but at different rates. Baseline performance shown in green.
2502.04173v1
2
https://arxiv.org/html/2…2.04172v1/x7.png
Fig. 2:Comparison of SMO and FNNLS[13]. In all instances of K (number of archetypes), the SMO updates converges to the optimal solution withinK2superscript𝐾2K^{2}italic_K start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPTiterations. The optimal solution for the differentK𝐾Kitalic_K’s are marked by the dotted black lines ide...
2502.04172v1
8
https://arxiv.org/html/2…3v1/pic/pic3.png
Figure 1:Consumption patterns of different entities can be similar (entity 1 and entity 2). Single-task learning techniques can only leverage information from the corresponding entity, while multi-task learning techniques use information from multiple entities.
2502.04163v1
1
https://arxiv.org/html/2…30_annotated.png
Figure 4.Mexico City, daily, 6:00–9:00–12:00: time-elapsed effective distances of OD pairs. Top 0.95 quantile per day. Legend shows the frequency of occurrence in days of the OD pairs. Circled: on 08-22-2019, OD pair (9g3qy,9g3qv) with effective distance 10.2.
2502.04162v1
8
https://arxiv.org/html/2…30_annotated.png
Figure 6.Mexico City, daily, 6:00–9:00–12:00: time-elapsed effective distances of globally-unconnected OD pairs. Top 0.95 quantile per day. Legend shows the frequency of occurrence in days of such OD pairs. Circled: on 09-23-2019, OD pair (9g3w6,9g3w0) with effective distance 2.16.
2502.04162v1
13
https://arxiv.org/html/2…2/figures/f1.jpg
Figure 1:Annotation Highlights.We capture multi-day recordings of unscripted activities.Centre-Top: Recipes are recorded with steps and their preparation temporally annotated, along with ingredient addition. Ingredients are weighed and nutrition recorded.Centre-Middle: Dense fine-grained narrations detailing what, how,...
2502.04144v1
1
https://arxiv.org/html/2…2.04144v1/x1.png
Figure 5:Frequency of verb clusters (top) and noun clusters (bottom) in narrated sentences by category, shown on a logarithmic scale.
2502.04144v1
9
https://arxiv.org/html/2…ionTrackerV7.jpg
Figure 6:Nutrition is monitored throughout recipes as ingredients are incorporated into dishes. Here we show Banana Bread Chocolate Chip Cookies. We annotate when ingredients are weighed, document their nutrition, and locate their adding time, giving us overall dish nutrition at each stage.
2502.04144v1
10
https://arxiv.org/html/2…ures/blender.jpg
Figure 7:Digital Twin: from point cloud (left), to surfaces (middle) and labelled fixtures (right). We show two moved objects (masks on top) at fixtures:cheeseandpan. Body poses from[85].
2502.04144v1
11
https://arxiv.org/html/2…2.04144v1/x2.png
Figure 10:VQA Question Prototypes. We show our 30 question prototypes by category alongside the number of questions. Outer bars indicate the distribution over input lengths for each question.
2502.04144v1
14
https://arxiv.org/html/2…e_hist_final.png
Figure 12:Effect of Input Length. Models struggle with questions of all video input lengths. s=second, m=minute, h=hour.
2502.04144v1
16
https://arxiv.org/html/2…ark_figure_2.png
Figure 14:Long-Term VOS. jaccard index𝒥𝒥\mathcal{J}caligraphic_J& contour accuracyℱℱ\mathcal{F}caligraphic_Fshow Cutie and SAM2 struggle with segmenting objects.
2502.04144v1
18
https://arxiv.org/html/2…framework_V4.png
Figure A3:Per ingredient, we temporally label weighing (if present) and adding to recipe. Weighing is OCRed and manually verified.
2502.04144v1
21
https://arxiv.org/html/2…2.04144v1/x5.png
Figure A9:Left: Common source→→\rightarrow→destination object movements. Right: Common take/put locations of the top-20 moving objects.
2502.04144v1
27
https://arxiv.org/html/2…ns_interface.png
Figure A10:Object associations interface. For each query track, a set of possible object names extracted from the narrations is suggested, we provide these names as an additional annotations. Importantly, all previous tracks are sorted by their distance. It is trivial for the annotator to confirm that the track on the ...
2502.04144v1
28
https://arxiv.org/html/2…r_dist_final.png
Figure A11:Prediction Bias of Models. Most models have a bias in answer, although it is different for each model.
2502.04144v1
29
https://arxiv.org/html/2…2.04143v1/x1.png
Figure 1:Schematic of the two-microphone method above a baffled porous layer, flush-mounted with a rigid backing.
2502.04143v1
1
https://arxiv.org/html/2…2.04143v1/x2.png
Figure 2:Experimental setup of the two-microphone method in the anechoic chamber at the Marcus Wallenberg Laboratory for Sound and Vibration in Stockholm.
2502.04143v1
2
https://arxiv.org/html/2…2.04143v1/x5.png
Figure 5:Histograms of the recorded error per sample for the two-microphone method and for the neural network predictions using the numerical test set.
2502.04143v1
5
https://arxiv.org/html/2…raph-image-2.png
Figure 3:Plot of one scenario of an infection from the diffusion SI-Equation. The Infection wave propagates over the nodes.
2502.04140v1
8
https://arxiv.org/html/2…on_diffusion.png
Figure 4:Plot of some advection-field on the left side and an exemplary plot of the solution of the advection-diffusion equation over 4 time-steps.
2502.04140v1
9
https://arxiv.org/html/2…SI_over_time.png
Figure 6:The dataset generated from the SI-diffusion equation over time. There are 400 lines plotted, each representing the values of an individual node over time.
2502.04140v1
11
https://arxiv.org/html/2…a_comparison.png
Figure 7:The SI-diffusion dataset(top) compared against the real-world German Influenza dataset(bottom). The step patterns in the Influenza dataset are presumably artefacts of the data collection/management pipeline.
2502.04140v1
12
https://arxiv.org/html/2…2.04140v1/x1.png
Figure 9:Illustration of the effect of additive noise sampled from the distribution𝒩⁢(0,0.01)𝒩00.01\mathcal{N}(0,0.01)caligraphic_N ( 0 , 0.01 )from task 2 and task 4. On the left is the time series from a random graph node; on the right is its respective noise-added time series. The plotted section is the first10001...
2502.04140v1
16
https://arxiv.org/html/2…k_1_variance.png
Figure 10:The RMSE over all samples from the test dataset. The less saturated colors represent the standard deviation for different initializations.
2502.04140v1
17
https://arxiv.org/html/2…_repetition2.png
Figure 13:Plots of the RMSEs of the presented models during the transfer-learning experiments. The arrows point from non-pre-trained to pre-trained models.
2502.04140v1
24
https://arxiv.org/html/2…y_plots_both.png
Figure 1:(Left)Comparison of Shannon entropy, Rényi entropy, and behavioral entropy (ours) and their effects on dataset generation, shown in PHATE plots, when used as an exploration objective.(Right)Performance comparison of an offline RL algorithm (CQL) for three downstream tasks on datasets generated using Shannon, b...
2502.04141v1
1
https://arxiv.org/html/2…elec_cividis.png
(a)Perception of normal distribution (𝒩⁢(0,1)𝒩01\mathcal{N}(0,1)caligraphic_N ( 0 , 1 )) under Prelec weighting function.Black dotted line shows standard normal density,f𝑓fitalic_f. Forα≈0𝛼0\alpha\approx 0italic_α ≈ 0the perceived density is uniform, indicatingover-weightingof uncertainty over entire support off𝑓f...
2502.04141v1
2
https://arxiv.org/html/2…elec_cividis.png
(a)Perception of normal distribution (𝒩⁢(0,1)𝒩01\mathcal{N}(0,1)caligraphic_N ( 0 , 1 )) under Prelec weighting function.Black dotted line shows standard normal density,f𝑓fitalic_f. Forα≈0𝛼0\alpha\approx 0italic_α ≈ 0the perceived density is uniform, indicatingover-weightingof uncertainty over entire support off𝑓f...
2502.04141v1
3
https://arxiv.org/html/2…gs/bernoulli.png
(b)Diversity of Shannon, Rényi and BE values as a function of Bernoulli trial parameterp𝑝pitalic_p.Behavioral entropyHBsuperscript𝐻𝐵H^{B}italic_H start_POSTSUPERSCRIPT italic_B end_POSTSUPERSCRIPTcaptures entire behavior spectrum from overvaluing uncertainty (light blue,α≈0𝛼0\alpha\approx 0italic_α ≈ 0) to highly u...
2502.04141v1
4
https://arxiv.org/html/2…lker_full_v2.png
(b)PHATE plots for Renyi for Walker. Coverage forq=3.0,5.0𝑞3.05.0q=3.0,5.0italic_q = 3.0 , 5.0similar toq=2.0𝑞2.0q=2.0italic_q = 2.0.
2502.04141v1
7
https://arxiv.org/html/2…on3M_SMM_RND.png
Figure 4:Comparison of offline RL performance over the entropy objectives used in dataset generation. Plots show mean and standard deviation over five seeds. Dotted line shows performance of RL policy trained online until approximate optimality.
2502.04141v1
8
https://arxiv.org/html/2…D_all_alphas.png
Figure 5:Offline RL results for allα𝛼\alphaitalic_αandq𝑞qitalic_qvalues evaluated. Initial trials showedq∈{2.0,3.0,5.0}𝑞2.03.05.0q\in\{2.0,3.0,5.0\}italic_q ∈ { 2.0 , 3.0 , 5.0 }led to performance no better (and usually worse) thanq=1.1𝑞1.1q=1.1italic_q = 1.1, so offline RL training for theseq𝑞qitalic_qvalues was ...
2502.04141v1
14
https://arxiv.org/html/2…200K_Shannon.png
Figure 6:Ablation result comparing the effect of performing 100K vs. 200K offline RL training steps on a 3M-element dataset generated using Shannon entropy as exploration objective. These results suggest that performing additional offline RL training has only a marginal effect on downstream task performance.
2502.04141v1
20
https://arxiv.org/html/2…reward_fn_BE.png
Figure 13:Visualization of the BE reward function equation26by varying the parameterα𝛼\alphaitalic_αwithβ𝛽\betaitalic_βconditioned according to (4) from(Suresh et al.,2024)withM=512𝑀512M=512italic_M = 512, denoting the representation dimensions. These visualizations highlight the diversity and variety of rewards tha...
2502.04141v1
49
https://arxiv.org/html/2…2.04139v1/x2.png
Figure 2:(a) The comparison of different query initialization methods.The FPS-based methods conduct farthest point sampling separately for each scene, placing more emphasis on positional information but lacking in aggregating content information. The learnable-based methods initialize a fixed number of queries for aggr...
2502.04139v1
2
https://arxiv.org/html/2…2.04139v1/x3.png
Figure 3:The overall framework of our method BFL.The Agent-Interpolation Initialization Module is meticulously crafted to synergize the strengths of FPS and learnable queries, producing object queries better suited for complex and dynamic environments. The Hierarchical Query Fusion Decoder is utilized to retain low ove...
2502.04139v1
3
https://arxiv.org/html/2…2.04139v1/x4.png
Figure 6:Visualization of instance segmentation results on ScanNetV2 validation set.The red boxes highlight the key regions.
2502.04139v1
5
https://arxiv.org/html/2…2.04139v1/x5.png
Figure 12:Parameter and runtime analysis of different methods on ScanNetV2 validation set.The runtime is measured on the same device.
2502.04139v1
8
https://arxiv.org/html/2…2.04139v1/x6.png
Figure 14:The architecture of the transformer decoder layer.The figure is taken from the main text of Maft.
2502.04139v1
9
https://arxiv.org/html/2…2.04139v1/x7.png
Figure 15:Additional Visual Comparison on ScanNetV2 validation set.The red boxes highlight the key regions.
2502.04139v1
10
https://arxiv.org/html/2…2.04139v1/x8.png
Figure 16:Visual comparisons between the baseline and our method across different decoder layers on ScanNetV2 validation set.The red boxes highlight the key regions.
2502.04139v1
11
https://arxiv.org/html/2…2.04139v1/x9.png
Figure 17:Visual comparisons between the baseline and our method across different decoder layers on ScanNetV2 validation set.The red boxes highlight the key regions.
2502.04139v1
12
https://arxiv.org/html/2….04139v1/x10.png
Figure 18:Visual comparisons between the baseline and our method across different decoder layers on ScanNetV2 validation set.The red boxes highlight the key regions.
2502.04139v1
13
https://arxiv.org/html/2…n-daughter-1.png
Figure 1:HowGPT-4o miniresponds to the same question when the order of choices is reversed. The calls were made on Wednesday, January 29th, at 11:54 a.m. EST.
2502.04134v1
1
https://arxiv.org/html/2…2.04131v1/x1.png
Figure 1:Geometric intuition behind the mechanism of SIM with data from the toy model. Panela)depicts a binary classification problem which illustrates how training data can be oriented along manifolds of the formΦ=g⁢(a,b)=a⁢bΦ𝑔𝑎𝑏𝑎𝑏\Phi=g(a,b)=abroman_Φ = italic_g ( italic_a , italic_b ) = italic_a italic_b. Panel...
2502.04131v1
1
https://arxiv.org/html/2…2.04131v1/x4.png
Figure 4:Binary classification task for time series from the batch reactor model. Displayed are 10 time series per class. Observational noise simulated is normally distributed with standard deviationσ=3𝜎3\sigma=3italic_σ = 3.
2502.04131v1
4
https://arxiv.org/html/2…2.04131v1/x5.png
Figure 5:Experiment 1 showing improved classification with partially observed batch reactor model due to SIM. Displayed are learning curves obtained from classifier training based on the fully observed (FO) dynamical model (dotted green) and the partially observed (PO) dynamical model, with and without application of S...
2502.04131v1
5
https://arxiv.org/html/2…2.04131v1/x6.png
Figure 6:Experiment 2 for the partially observed batch reactor model showing that SIM is robust to observational noise. Displayed are generalization error and relative number of support vectors, each as a function of the observational noise. Classification performance is compared for the partially observed (PO) dynamic...
2502.04131v1
6
https://arxiv.org/html/2…2.04131v1/x7.png
Figure 7:Experiment 3 for the partially observed batch reactor model showing SIM is robust to changes in regularity and sparsity of the observed time series data. Displayed are the learning curves obtained from classifier training for the partially observed (PO) dynamical model, with and without application of SIM (mar...
2502.04131v1
7