{"noiseShare":0.5133,"topics":[{"claims":[509,610,856,971,1087,1347,1725,2397,3655],"deltaPp":1.7778,"id":0,"keywords":["state-of-the-art","inference","outperforms","benchmark","experiments"],"label":"State-of-the-Art Performance","labelFull":"State-of-the-Art Performance on Standard Benchmarks","papers":11461,"papersByYear":[453,549,750,871,963,1190,1541,2100,3044],"rel":0.257,"reviewed":true,"share":[6.9181,6.6586,7.1449,7.3224,7.3635,7.6983,7.2514,7.7809,8.6959],"size":13157,"slope":0.001844},{"claims":[414,686,1084,1125,1272,1710,1933,2239,2324],"deltaPp":-0.0111,"id":1,"keywords":["graph","node","networks","neural","classification"],"label":"Efficient Graph Neural Networks","labelFull":"Efficient and Scalable Graph Neural Network Architectures","papers":5550,"papersByYear":[205,309,483,513,523,676,798,951,1092],"rel":-0.0035,"reviewed":true,"share":[3.1307,3.7477,4.6013,4.3127,3.9991,4.3731,3.7551,3.5237,3.1196],"size":12787,"slope":-0.000391},{"claims":[511,821,1150,1217,1328,1373,1544,1877,2193],"deltaPp":-0.5587,"id":2,"keywords":["adversarial","attacks","backdoor","robustness","defense"],"label":"Adversarial Robustness","labelFull":"Adversarial Robustness: Attacks, Defenses, and Training","papers":5780,"papersByYear":[259,400,527,562,572,600,721,950,1189],"rel":-0.1412,"reviewed":true,"share":[3.9554,4.8514,5.0205,4.7247,4.3738,3.8815,3.3928,3.52,3.3967],"size":12014,"slope":-0.001721},{"claims":[189,266,361,406,567,712,2032,2928,2723],"deltaPp":3.6578,"id":3,"keywords":["languages","multilingual","fine-tuning","evaluation","framework"],"label":"LLM Benchmarking","labelFull":"Evaluation and Benchmarking of Large Language Models","papers":6960,"papersByYear":[126,181,243,286,388,485,1366,1931,1954],"rel":1.9008,"reviewed":true,"share":[1.9243,2.1953,2.3149,2.4044,2.9668,3.1375,6.4279,7.1548,5.5821],"size":10184,"slope":0.006411},{"claims":[371,611,744,940,1060,1303,1426,1493,1406],"deltaPp":-0.663,"id":4,"keywords":["segmentation","semantic","instance","medical","object"],"label":"Image Segmentation","labelFull":"Image Segmentation: Methods, Benchmarks, and Applications","papers":4998,"papersByYear":[194,317,404,492,555,692,753,786,805],"rel":-0.2238,"reviewed":true,"share":[2.9627,3.8448,3.8487,4.1362,4.2438,4.4766,3.5434,2.9123,2.2997],"size":9354,"slope":-0.000953},{"claims":[89,140,194,200,271,443,914,2109,4684],"deltaPp":7.0438,"id":5,"keywords":["reasoning","mathematical","math","chain-of-thought","benchmarks"],"label":"Efficient Chain-of-Thought Reasoning","labelFull":"Efficient Chain-of-Thought Reasoning for LLMs","papers":5236,"papersByYear":[53,83,117,115,158,261,519,1181,2749],"rel":8.7025,"reviewed":true,"share":[0.8094,1.0067,1.1146,0.9668,1.2081,1.6884,2.4422,4.3759,7.8532],"size":9044,"slope":0.006943},{"claims":[240,308,358,565,665,850,1509,1848,2249],"deltaPp":0.7726,"id":6,"keywords":["speech","audio","music","acoustic","recognition"],"label":"Audio-Language Models","labelFull":"Audio-Language Models and Pretraining for Audio Understanding","papers":3029,"papersByYear":[97,126,138,210,242,315,499,613,789],"rel":0.5215,"reviewed":true,"share":[1.4814,1.5282,1.3147,1.7654,1.8504,2.0378,2.3481,2.2713,2.254],"size":8592,"slope":0.001276},{"claims":[86,132,237,336,454,834,1235,2158,3028],"deltaPp":2.5557,"id":7,"keywords":["molecular","protein","quantum","prediction","design"],"label":"Molecular Machine Learning","labelFull":"Machine Learning for Molecular and Protein Design","papers":3351,"papersByYear":[56,64,110,156,193,335,460,783,1194],"rel":2.9884,"reviewed":true,"share":[0.8552,0.7762,1.0479,1.3115,1.4758,2.1672,2.1646,2.9012,3.4109],"size":8500,"slope":0.003281},{"claims":[514,559,702,702,758,666,1004,1336,1538],"deltaPp":-0.1352,"id":8,"keywords":["question","answering","qa","reading","comprehension"],"label":"Question Answering","labelFull":"Question Answering across Multiple-Choice, Open-Domain, and Visual","papers":4188,"papersByYear":[212,230,315,319,342,338,538,808,1086],"rel":-0.0418,"reviewed":true,"share":[3.2376,2.7896,3.0009,2.6818,2.6151,2.1866,2.5316,2.9938,3.1024],"size":7779,"slope":-0.000227},{"claims":[316,459,671,898,1020,1098,1075,1073,840],"deltaPp":-1.6058,"id":9,"keywords":["label","contrastive","loss","learning","class"],"label":"Learning with Noisy Labels","labelFull":"Robust Classification under Noisy Labels","papers":4952,"papersByYear":[220,318,436,605,669,699,701,690,614],"rel":-0.4779,"reviewed":true,"share":[3.3598,3.8569,4.1536,5.0862,5.1155,4.5219,3.2987,2.5566,1.754],"size":7450,"slope":-0.0021},{"claims":[378,474,587,771,850,931,1157,1264,1025],"deltaPp":-0.9756,"id":10,"keywords":["pose","motion","human","estimation","3d"],"label":"3D Human Pose Estimation","labelFull":"3D Human Pose Estimation from Images and Video","papers":3655,"papersByYear":[172,221,264,337,391,463,586,643,578],"rel":-0.3714,"reviewed":true,"share":[2.6268,2.6804,2.515,2.8331,2.9898,2.9952,2.7575,2.3825,1.6512],"size":7437,"slope":-0.000691},{"claims":[37,74,95,138,278,550,989,1860,2916],"deltaPp":5.0449,"id":11,"keywords":["vision-language","visual","reasoning","tokens","text"],"label":"Vision-Language Reasoning","labelFull":"Improving and Assessing Reasoning in Vision-Language Models","papers":4595,"papersByYear":[29,57,72,101,191,386,660,1178,1921],"rel":11.3906,"reviewed":true,"share":[0.4429,0.6913,0.6859,0.8491,1.4605,2.4971,3.1057,4.3647,5.4878],"size":6937,"slope":0.006281},{"claims":[361,561,713,728,716,769,918,1003,1052],"deltaPp":-0.7826,"id":12,"keywords":["dialogue","conversational","response","user","generation"],"label":"Dialogue Systems","labelFull":"Dialogue Systems: Modeling, Generation, and Understanding","papers":3043,"papersByYear":[156,226,279,296,285,323,430,488,560],"rel":-0.3285,"reviewed":true,"share":[2.3824,2.7411,2.6579,2.4884,2.1792,2.0895,2.0234,1.8081,1.5998],"size":6821,"slope":-0.001266},{"claims":[545,682,774,731,725,777,715,784,893],"deltaPp":-3.1874,"id":13,"keywords":["networks","neural","relu","activation","deep"],"label":"Layerwise Training","labelFull":"Layerwise Training for Efficient Deep Neural Networks","papers":4220,"papersByYear":[326,405,458,477,453,487,478,509,627],"rel":-0.6402,"reviewed":true,"share":[4.9786,4.9121,4.3632,4.0101,3.4638,3.1505,2.2493,1.886,1.7912],"size":6626,"slope":-0.004486},{"claims":[553,606,736,727,832,673,876,756,562],"deltaPp":-2.9039,"id":14,"keywords":["entity","relation","extraction","knowledge graph","named"],"label":"Entity Linking & Relation Extraction","labelFull":"Entity Linking and Relation Extraction for Knowledge Graphs","papers":3044,"papersByYear":[253,286,337,328,382,321,418,383,336],"rel":-0.7516,"reviewed":true,"share":[3.8638,3.4688,3.2104,2.7575,2.9209,2.0766,1.967,1.4191,0.9599],"size":6321,"slope":-0.003489},{"claims":[432,469,557,576,627,716,841,801,1007],"deltaPp":-1.9438,"id":15,"keywords":["convergence","gradient","stochastic","descent","convex"],"label":"Accelerated Optimization Methods","labelFull":"Accelerated Optimization Algorithms with Fast Convergence","papers":3525,"papersByYear":[247,277,307,328,357,416,468,485,640],"rel":-0.5153,"reviewed":true,"share":[3.7721,3.3596,2.9246,2.7575,2.7298,2.6912,2.2022,1.797,1.8283],"size":6026,"slope":-0.002329},{"claims":[266,318,423,576,651,738,841,826,1342],"deltaPp":0.2556,"id":16,"keywords":["reinforcement","learning","offline","policy","imitation"],"label":"Offline Reinforcement Learning","labelFull":"Offline Reinforcement Learning Algorithms and Benchmarks","papers":3986,"papersByYear":[172,213,285,353,417,448,529,560,1009],"rel":0.0973,"reviewed":true,"share":[2.6268,2.5834,2.7151,2.9676,3.1886,2.8982,2.4893,2.0749,2.8824],"size":5981,"slope":-0.000171},{"claims":[11,17,21,51,121,552,1198,1699,2266],"deltaPp":3.703,"id":17,"keywords":["diffusion","denoising","sampling","process","latent"],"label":"Efficient Diffusion Modeling","labelFull":"Efficient Training and Sampling Techniques for Diffusion Models","papers":3454,"papersByYear":[8,14,16,34,65,306,695,977,1339],"rel":30.3028,"reviewed":true,"share":[0.1222,0.1698,0.1524,0.2858,0.497,1.9796,3.2704,3.62,3.8252],"size":5936,"slope":0.005515},{"claims":[69,92,126,168,242,364,762,1590,2251],"deltaPp":3.4531,"id":18,"keywords":["multimodal","modality","fusion","reasoning","alignment"],"label":"Multimodal Foundation Models","labelFull":"Multimodal Foundation Models: Alignment and Evaluation","papers":3586,"papersByYear":[44,66,88,114,157,213,477,983,1444],"rel":5.1385,"reviewed":true,"share":[0.672,0.8005,0.8383,0.9584,1.2005,1.3779,2.2446,3.6422,4.1251],"size":5664,"slope":0.004262},{"claims":[82,151,201,193,177,220,501,1115,2987],"deltaPp":4.0681,"id":19,"keywords":["agents","planning","llm","systems","evaluation"],"label":"LLM Agents","labelFull":"LLM Agents for Planning, Tool Use, and Evaluation","papers":3285,"papersByYear":[52,91,120,132,123,147,297,621,1702],"rel":5.1229,"reviewed":true,"share":[0.7941,1.1037,1.1432,1.1097,0.9405,0.951,1.3976,2.3009,4.8622],"size":5627,"slope":0.003369},{"claims":[90,123,154,238,328,396,725,1362,1828],"deltaPp":2.3214,"id":20,"keywords":["retrieval","query","generation","information","accuracy"],"label":"Retrieval-Augmented Generation","labelFull":"Retrieval-Augmented Generation for Large Language Models","papers":3201,"papersByYear":[59,83,107,149,209,256,431,779,1128],"rel":2.5765,"reviewed":true,"share":[0.901,1.0067,1.0193,1.2526,1.5981,1.6561,2.0281,2.8864,3.2224],"size":5244,"slope":0.002891},{"claims":[74,135,271,343,372,490,612,936,1656],"deltaPp":2.1392,"id":21,"keywords":["flow","pde","dynamics","equations","langevin"],"label":"Neural PDE Solvers","labelFull":"Neural PDE Solvers and Physics-Informed Neural Networks","papers":2726,"papersByYear":[41,80,155,187,212,262,329,492,968],"rel":3.4167,"reviewed":true,"share":[0.6261,0.9703,1.4766,1.5721,1.621,1.6949,1.5482,1.823,2.7653],"size":4889,"slope":0.001897},{"claims":[516,518,669,637,596,452,588,495,407],"deltaPp":-3.049,"id":22,"keywords":["translation","machine","neural","quality","english"],"label":"Neural Machine Translation","labelFull":"Neural Machine Translation: Low-Resource Domain Adaptation","papers":2212,"papersByYear":[235,240,297,293,255,209,260,234,189],"rel":-0.8496,"reviewed":true,"share":[3.5889,2.9109,2.8294,2.4632,1.9498,1.3521,1.2235,0.867,0.5399],"size":4878,"slope":-0.003775},{"claims":[111,232,390,526,623,749,802,769,599],"deltaPp":0.0283,"id":23,"keywords":["point","cloud","3d","lidar","object"],"label":"3D Point Cloud Understanding","labelFull":"Deep Learning for 3D Point Cloud Understanding","papers":2459,"papersByYear":[61,109,196,247,302,398,410,400,336],"rel":0.0304,"reviewed":true,"share":[0.9316,1.322,1.8672,2.0765,2.3092,2.5747,1.9293,1.4821,0.9599],"size":4801,"slope":0.000203},{"claims":[87,141,162,293,375,678,944,1128,966],"deltaPp":0.9229,"id":24,"keywords":["rendering","gaussian","radiance","splatting","view synthesis"],"label":"3D Gaussian Splatting","labelFull":"3D Gaussian Splatting for Real-Time Novel View Synthesis","papers":2693,"papersByYear":[49,73,90,156,205,376,537,622,585],"rel":1.2333,"reviewed":true,"share":[0.7483,0.8854,0.8574,1.3115,1.5675,2.4324,2.5269,2.3046,1.6712],"size":4774,"slope":0.002068},{"claims":[432,406,504,492,553,625,586,723,372],"deltaPp":-1.9875,"id":25,"keywords":["face","facial","recognition","expression","identity"],"label":"Identity-Preserving Face Generation","labelFull":"Identity-Preserving Face Generation and Manipulation","papers":1859,"papersByYear":[164,157,196,182,213,220,240,306,181],"rel":-0.7935,"reviewed":true,"share":[2.5046,1.9042,1.8672,1.5301,1.6287,1.4232,1.1294,1.1338,0.5171],"size":4693,"slope":-0.001974},{"claims":[189,236,362,398,428,586,731,734,832],"deltaPp":0.1344,"id":26,"keywords":["super-resolution","restoration","denoising","image","reconstruction"],"label":"Image Restoration","labelFull":"Deep Learning for Image Restoration and Super-Resolution","papers":2634,"papersByYear":[101,126,184,201,233,325,416,461,587],"rel":0.0871,"reviewed":true,"share":[1.5425,1.5282,1.7529,1.6898,1.7816,2.1025,1.9576,1.7081,1.6769],"size":4496,"slope":0.000317},{"claims":[117,145,259,377,358,581,729,799,1028],"deltaPp":0.6074,"id":27,"keywords":["causal","counterfactual","treatment","effect","variables"],"label":"Causal Inference Benchmarking","labelFull":"Benchmarking Causal Discovery and Effect Estimation","papers":2105,"papersByYear":[59,71,120,180,176,270,329,372,528],"rel":0.6741,"reviewed":true,"share":[0.901,0.8611,1.1432,1.5132,1.3458,1.7467,1.5482,1.3783,1.5084],"size":4393,"slope":0.000837},{"claims":[15,29,72,105,192,279,473,1029,2048],"deltaPp":4.0199,"id":28,"keywords":["token","attention","prediction","information"],"label":"Dynamic Token Pruning","labelFull":"Dynamic Token Pruning for Efficient Transformer Inference","papers":3086,"papersByYear":[14,23,56,85,141,213,336,736,1482],"rel":18.8022,"reviewed":true,"share":[0.2138,0.279,0.5335,0.7146,1.0781,1.3779,1.5811,2.727,4.2337],"size":4242,"slope":0.004364},{"claims":[41,99,218,308,421,532,610,773,1099],"deltaPp":1.7332,"id":29,"keywords":["distillation","knowledge","teacher","student","loss"],"label":"Knowledge Distillation","labelFull":"Knowledge Distillation for Efficient Model Compression","papers":2521,"papersByYear":[24,57,118,178,243,315,377,474,735],"rel":4.7291,"reviewed":true,"share":[0.3665,0.6913,1.1241,1.4964,1.8581,2.0378,1.774,1.7563,2.0997],"size":4101,"slope":0.001995},{"claims":[79,102,136,214,207,342,600,915,1396],"deltaPp":1.0991,"id":30,"keywords":["time series","forecasting","multivariate","temporal","real-world"],"label":"Time Series Forecasting","labelFull":"Deep Learning for Time Series Forecasting","papers":1847,"papersByYear":[47,59,79,113,102,162,252,397,636],"rel":1.5312,"reviewed":true,"share":[0.7178,0.7156,0.7526,0.95,0.7799,1.048,1.1858,1.471,1.8169],"size":3991,"slope":0.001271},{"claims":[178,293,408,463,535,500,487,395,415],"deltaPp":-0.6873,"id":31,"keywords":["detection","object","detectors","coco","state-of-the-art"],"label":"Object Detection","labelFull":"State-of-the-Art Object Detection Methods and Benchmarks","papers":2154,"papersByYear":[100,159,211,244,296,296,287,267,294],"rel":-0.45,"reviewed":true,"share":[1.5272,1.9284,2.0101,2.0513,2.2633,1.9149,1.3505,0.9893,0.8399],"size":3674,"slope":-0.00117},{"claims":[126,169,242,397,506,559,547,526,508],"deltaPp":-0.3293,"id":32,"keywords":["vision","transformers","visual","image","classification"],"label":"Vision Transformers","labelFull":"Vision Transformers: Pretraining, Efficiency, and Robustness","papers":2811,"papersByYear":[102,146,207,299,364,432,421,410,430],"rel":-0.2114,"reviewed":true,"share":[1.5577,1.7708,1.972,2.5137,2.7833,2.7947,1.9811,1.5191,1.2284],"size":3580,"slope":-0.000296},{"claims":[124,232,285,385,448,516,515,546,485],"deltaPp":0.065,"id":33,"keywords":["few-shot","zero-shot","classification","learning","generalization"],"label":"Few-Shot and Zero-Shot Learning","labelFull":"Few-Shot and Zero-Shot Learning for Classification","papers":2445,"papersByYear":[73,123,170,235,286,363,369,413,413],"rel":0.0583,"reviewed":true,"share":[1.1148,1.4918,1.6195,1.9756,2.1869,2.3483,1.7364,1.5303,1.1798],"size":3536,"slope":0.000164},{"claims":[92,158,184,282,274,272,418,714,1109],"deltaPp":0.9115,"id":34,"keywords":["attention","heads","linear","sparse","mechanism"],"label":"Efficient Attention","labelFull":"Efficient Attention Mechanisms for Transformers","papers":2311,"papersByYear":[72,100,127,180,184,188,301,455,704],"rel":0.8289,"reviewed":true,"share":[1.0996,1.2129,1.2099,1.5132,1.4069,1.2162,1.4164,1.6859,2.0111],"size":3503,"slope":0.000863},{"claims":[140,193,254,249,307,317,429,632,977],"deltaPp":0.6271,"id":35,"keywords":["quantization","compression","quantized","post-training","weight"],"label":"Low-Precision Quantization","labelFull":"Low-Precision Quantization and QAT for Efficient Models","papers":1890,"papersByYear":[65,94,122,125,152,169,245,351,567],"rel":0.6317,"reviewed":true,"share":[0.9927,1.1401,1.1622,1.0509,1.1623,1.0933,1.1529,1.3005,1.6198],"size":3498,"slope":0.000502},{"claims":[270,297,331,334,409,395,353,503,382],"deltaPp":-1.1039,"id":36,"keywords":["depth","stereo","estimation","monocular","disparity"],"label":"Monocular Depth Estimation","labelFull":"Monocular Depth Estimation from Single Images","papers":1569,"papersByYear":[114,124,152,139,185,194,183,255,223],"rel":-0.6341,"reviewed":true,"share":[1.741,1.5039,1.448,1.1686,1.4146,1.255,0.8611,0.9448,0.6371],"size":3274,"slope":-0.001197},{"claims":[0,3,2,10,36,347,974,1077,804],"deltaPp":1.7855,"id":37,"keywords":["diffusion","text-to-image","video","generation","latent"],"label":"Diffusion-based Image/Video Generation","labelFull":"Diffusion-based Image/Video Generation, Control, Efficiency","papers":2393,"papersByYear":[0,3,2,8,23,242,679,811,625],"rel":null,"reviewed":true,"share":[0.0,0.0364,0.0191,0.0673,0.1759,1.5655,3.1951,3.0049,1.7855],"size":3253,"slope":0.003983},{"claims":[69,115,174,206,243,307,376,663,1095],"deltaPp":1.3114,"id":38,"keywords":["navigation","manipulation","robotic","locomotion","vision-and-language"],"label":"Embodied AI","labelFull":"Embodied AI for Navigation, Manipulation, and Locomotion","papers":2088,"papersByYear":[49,83,104,143,152,194,244,398,721],"rel":1.7525,"reviewed":true,"share":[0.7483,1.0067,0.9908,1.2022,1.1623,1.255,1.1482,1.4747,2.0597],"size":3248,"slope":0.00117},{"claims":[282,351,370,376,323,369,344,381,444],"deltaPp":-1.6496,"id":39,"keywords":["variational","posterior","gaussian","bayesian","inference"],"label":"Variational Bayesian Neural Networks","labelFull":"Variational Inference for Bayesian Neural Networks","papers":2137,"papersByYear":[169,227,229,252,202,238,235,259,326],"rel":-0.6392,"reviewed":true,"share":[2.5809,2.7532,2.1816,2.1185,1.5446,1.5397,1.1058,0.9597,0.9313],"size":3240,"slope":-0.002452},{"claims":[270,297,322,393,497,338,478,381,245],"deltaPp":-1.2457,"id":40,"keywords":["summarization","abstractive","extractive","document","evaluation"],"label":"Neural Summarization","labelFull":"Neural Summarization: Models, Benchmarks, and Evaluation","papers":1452,"papersByYear":[110,125,138,171,193,144,224,195,152],"rel":-0.7415,"reviewed":true,"share":[1.6799,1.5161,1.3147,1.4376,1.4758,0.9316,1.0541,0.7225,0.4342],"size":3221,"slope":-0.001398},{"claims":[192,291,399,446,424,421,333,240,293],"deltaPp":-1.6674,"id":41,"keywords":["imagenet","cifar-10","cifar-100","accuracy","top-1"],"label":"Image Classification Benchmarks","labelFull":"State-of-the-Art Image Classification on ImageNet and CIFAR","papers":2612,"papersByYear":[158,246,332,377,361,356,293,228,261],"rel":-0.691,"reviewed":true,"share":[2.413,2.9836,3.1628,3.1694,2.7604,2.303,1.3788,0.8448,0.7456],"size":3039,"slope":-0.00292},{"claims":[87,141,181,267,299,410,413,458,574],"deltaPp":0.21,"id":42,"keywords":["privacy","private","differential","guarantees","utility"],"label":"Differentially Private Machine Learning","labelFull":"Differential Privacy for Machine Learning: Theory & Practice","papers":1310,"papersByYear":[39,58,73,116,122,186,198,236,282],"rel":0.3526,"reviewed":true,"share":[0.5956,0.7035,0.6954,0.9752,0.9329,1.2033,0.9317,0.8744,0.8056],"size":2830,"slope":0.000342},{"claims":[114,170,223,209,239,222,410,632,600],"deltaPp":0.0837,"id":43,"keywords":["driving","autonomous","traffic","lane","vehicle"],"label":"End-to-End Autonomous Driving","labelFull":"End-to-End Autonomous Driving: Datasets, Models, Evaluation","papers":1396,"papersByYear":[54,71,106,104,118,120,198,307,318],"rel":0.1015,"reviewed":true,"share":[0.8247,0.8611,1.0098,0.8743,0.9023,0.7763,0.9317,1.1375,0.9084],"size":2819,"slope":0.000152},{"claims":[112,184,259,329,292,344,326,392,483],"deltaPp":-0.2295,"id":44,"keywords":["games","equilibrium","nash","equilibria","zero-sum"],"label":"Game Equilibrium Computation","labelFull":"Algorithms for Nash, Correlated, and Stackelberg Equilibria","papers":1147,"papersByYear":[53,73,102,137,123,140,147,169,203],"rel":-0.2835,"reviewed":true,"share":[0.8094,0.8854,0.9717,1.1517,0.9405,0.9057,0.6917,0.6262,0.5799],"size":2721,"slope":-0.000417},{"claims":[83,98,146,242,304,294,401,497,561],"deltaPp":0.2821,"id":45,"keywords":["explanations","claim","legal","rationales","evidence"],"label":"Rationale Generation","labelFull":"Rationale Generation for Explainable AI Predictions","papers":1466,"papersByYear":[47,54,83,130,154,168,215,265,350],"rel":0.393,"reviewed":true,"share":[0.7178,0.6549,0.7907,1.0929,1.1776,1.0868,1.0117,0.9819,0.9999],"size":2626,"slope":0.000424},{"claims":[107,181,230,335,317,383,344,369,320],"deltaPp":-0.3095,"id":46,"keywords":["bandits","regret","multi-armed","contextual","linear"],"label":"Bandit Regret Bounds","labelFull":"Regret Bounds for Multi-Armed, Linear and Contextual Bandits","papers":1272,"papersByYear":[53,89,114,155,157,192,155,182,175],"rel":-0.3824,"reviewed":true,"share":[0.8094,1.0794,1.086,1.3031,1.2005,1.2421,0.7294,0.6743,0.4999],"size":2586,"slope":-0.000538},{"claims":[48,63,145,154,231,285,338,557,763],"deltaPp":1.0625,"id":47,"keywords":["continual","forgetting","unlearning","curriculum","catastrophic"],"label":"Continual Learning","labelFull":"Continual Learning and Catastrophic Forgetting Mitigation","papers":1759,"papersByYear":[32,46,108,103,146,183,219,379,543],"rel":2.1741,"reviewed":true,"share":[0.4887,0.5579,1.0289,0.8659,1.1164,1.1839,1.0305,1.4043,1.5512],"size":2584,"slope":0.001185},{"claims":[176,248,184,199,185,275,360,496,443],"deltaPp":-0.4372,"id":48,"keywords":["captioning","video","image","understanding","visual"],"label":"Visual Captioning","labelFull":"Image and Video Captioning with Vision-Language Models","papers":1429,"papersByYear":[84,106,86,101,98,149,204,305,296],"rel":-0.3408,"reviewed":true,"share":[1.2828,1.2856,0.8193,0.8491,0.7494,0.9639,0.96,1.1301,0.8456],"size":2566,"slope":-0.000303},{"claims":[211,250,302,320,325,292,316,288,220],"deltaPp":-1.2095,"id":49,"keywords":["action","recognition","temporal","activity","localization"],"label":"Action Recognition & Detection","labelFull":"Video Action Recognition and Detection","papers":1380,"papersByYear":[108,133,149,164,168,162,174,168,154],"rel":-0.7333,"reviewed":true,"share":[1.6494,1.6131,1.4195,1.3787,1.2846,1.048,0.8188,0.6225,0.4399],"size":2524,"slope":-0.001557},{"claims":[91,119,175,210,194,300,315,498,614],"deltaPp":0.13,"id":50,"keywords":["search","combinatorial","solutions","optimization","pareto"],"label":"Multi-Objective Combinatorial Optimization","labelFull":"Multi-Objective Combinatorial Optimization Methods","papers":1549,"papersByYear":[65,79,109,126,134,172,183,288,393],"rel":0.131,"reviewed":true,"share":[0.9927,0.9582,1.0384,1.0593,1.0246,1.1127,0.8611,1.0671,1.1227],"size":2516,"slope":9.1e-05},{"claims":[110,163,163,205,150,202,301,542,652],"deltaPp":0.0378,"id":51,"keywords":["recommendation","user","item","collaborative","sequential"],"label":"Recommendation Systems","labelFull":"Collaborative and Sequential Recommendation Systems","papers":1252,"papersByYear":[63,72,90,97,77,119,138,246,350],"rel":0.0393,"reviewed":true,"share":[0.9621,0.8733,0.8574,0.8155,0.5888,0.7698,0.6494,0.9115,0.9999],"size":2488,"slope":-3.3e-05},{"claims":[84,131,234,272,291,318,335,332,461],"deltaPp":0.0297,"id":52,"keywords":["regret","bound","algorithm","online","sqrt"],"label":"Online Regret Minimization","labelFull":"Online Regret Minimization and Sublinear Regret Bounds","papers":1449,"papersByYear":[48,80,141,161,173,201,192,186,267],"rel":0.0405,"reviewed":true,"share":[0.733,0.9703,1.3432,1.3535,1.3228,1.3003,0.9035,0.6892,0.7627],"size":2458,"slope":-0.000276},{"claims":[410,413,356,271,277,162,258,210,100],"deltaPp":-3.1045,"id":53,"keywords":["word","embeddings","sense","sentence","morphological"],"label":"Embedding Evaluation","labelFull":"Evaluation and Benchmarking of Word and Sentence Embeddings","papers":1424,"papersByYear":[219,233,205,145,156,94,158,130,84],"rel":-0.9282,"reviewed":true,"share":[3.3445,2.826,1.9529,1.219,1.1928,0.6081,0.7435,0.4817,0.24],"size":2457,"slope":-0.003747},{"claims":[1,8,63,162,296,442,522,479,401],"deltaPp":0.656,"id":54,"keywords":["federated","learning","client","personalized","communication"],"label":"Federated Learning","labelFull":"Federated Learning: Heterogeneity and Communication","papers":1194,"papersByYear":[1,3,30,76,134,216,252,247,235],"rel":42.8758,"reviewed":true,"share":[0.0153,0.0364,0.2858,0.6389,1.0246,1.3973,1.1858,0.9152,0.6713],"size":2374,"slope":0.001303},{"claims":[122,182,261,316,323,344,320,278,222],"deltaPp":-0.5958,"id":55,"keywords":["domain","adaptation","source","target","generalization"],"label":"Domain Adaptation","labelFull":"Domain Adaptation: Unsupervised, Multi-Source, Open-Set","papers":1559,"papersByYear":[71,110,154,185,213,224,226,205,171],"rel":-0.5495,"reviewed":true,"share":[1.0843,1.3341,1.4671,1.5553,1.6287,1.4491,1.0635,0.7596,0.4885],"size":2368,"slope":-0.000837},{"claims":[23,27,36,51,108,187,472,644,820],"deltaPp":1.3552,"id":56,"keywords":["editing","generation","text-to-image","video","quality"],"label":"Text-to-Image Generation & Editing","labelFull":"Unified Text-to-Image Generation and Image Editing Techniques","papers":1769,"papersByYear":[22,23,32,38,76,145,342,499,592],"rel":4.0333,"reviewed":true,"share":[0.336,0.279,0.3048,0.3195,0.5811,0.938,1.6093,1.8489,1.6912],"size":2368,"slope":0.002226},{"claims":[120,151,180,297,319,264,258,367,385],"deltaPp":-0.3449,"id":57,"keywords":["video","frames","temporal","prediction","features"],"label":"Efficient Video Transformers","labelFull":"Efficient Video Transformers for Representation Learning","papers":1714,"papersByYear":[83,107,134,196,212,192,194,273,323],"rel":-0.2721,"reviewed":true,"share":[1.2676,1.2978,1.2766,1.6478,1.621,1.2421,0.9129,1.0115,0.9227],"size":2341,"slope":-0.000562},{"claims":[82,129,184,245,239,280,316,286,564],"deltaPp":0.2444,"id":58,"keywords":["policy","markov decision","decision processes","gradient","mdps"],"label":"Off-Policy Policy Gradient","labelFull":"Off-Policy Policy Gradient for Sample-Efficient Reinforcement Learning","papers":1605,"papersByYear":[62,87,127,168,155,183,207,199,417],"rel":0.2581,"reviewed":true,"share":[0.9469,1.0552,1.2099,1.4124,1.1852,1.1839,0.9741,0.7373,1.1913],"size":2325,"slope":-0.000113},{"claims":[76,68,125,138,187,233,387,422,601],"deltaPp":0.3074,"id":59,"keywords":["spiking","neural","networks","neurons","activity"],"label":"Spiking Neural Network Training","labelFull":"Efficient Training Methods for Spiking Neural Networks","papers":982,"papersByYear":[30,38,49,57,76,109,164,191,268],"rel":0.6709,"reviewed":true,"share":[0.4582,0.4609,0.4668,0.4792,0.5811,0.7051,0.7717,0.7077,0.7656],"size":2237,"slope":0.000468},{"claims":[31,64,58,82,83,143,356,530,873],"deltaPp":1.4962,"id":60,"keywords":["code","instruction","verification","test","generation"],"label":"Execution-Guided Code Generation","labelFull":"Execution-Guided Code Generation for Large Language Models","papers":1486,"papersByYear":[21,36,34,46,48,84,225,356,636],"rel":4.6654,"reviewed":true,"share":[0.3207,0.4366,0.3239,0.3867,0.367,0.5434,1.0588,1.3191,1.8169],"size":2220,"slope":0.00171},{"claims":[91,132,212,207,242,281,325,282,241],"deltaPp":-0.3342,"id":61,"keywords":["shape","surface","mesh","signed distance","3d"],"label":"Neural 3D Surface Reconstruction","labelFull":"Neural 3D Surface Reconstruction and Mesh Generation","papers":1218,"papersByYear":[52,71,131,127,142,172,188,174,161],"rel":-0.4209,"reviewed":true,"share":[0.7941,0.8611,1.248,1.0677,1.0858,1.1127,0.8847,0.6447,0.4599],"size":2013,"slope":-0.000445},{"claims":[106,115,113,148,122,169,288,440,456],"deltaPp":0.1264,"id":62,"keywords":["visual question","answering","video question","pairs","image"],"label":"Visual Question Answering","labelFull":"Visual Question Answering, Including Knowledge-Based Variants","papers":1299,"papersByYear":[60,60,62,79,79,96,179,319,365],"rel":0.1379,"reviewed":true,"share":[0.9163,0.7277,0.5906,0.6641,0.6041,0.621,0.8423,1.182,1.0427],"size":1957,"slope":0.000388},{"claims":[267,239,269,254,213,120,247,211,135],"deltaPp":-1.6298,"id":63,"keywords":["sentiment","aspect","analysis","opinion","classification"],"label":"Aspect-Based Sentiment Analysis","labelFull":"Aspect-Based Sentiment Analysis (ABSA)","papers":898,"papersByYear":[120,117,116,117,92,63,108,94,71],"rel":-0.8893,"reviewed":true,"share":[1.8326,1.419,1.1051,0.9836,0.7035,0.4076,0.5082,0.3483,0.2028],"size":1955,"slope":-0.001917},{"claims":[29,38,87,117,126,152,303,446,654],"deltaPp":0.997,"id":64,"keywords":["decoding","transformer","speculative","layers"],"label":"Efficient Transformers","labelFull":"Efficient Transformer Architectures and Training Techniques","papers":1458,"papersByYear":[26,35,62,96,102,112,225,312,488],"rel":2.5107,"reviewed":true,"share":[0.3971,0.4245,0.5906,0.8071,0.7799,0.7245,1.0588,1.156,1.3941],"size":1952,"slope":0.001173},{"claims":[2,1,5,15,134,237,422,532,554],"deltaPp":1.2493,"id":65,"keywords":["prompt","tuning","optimization","learning"],"label":"Prompt Optimization","labelFull":"Automated Prompt Optimization for Large Language Models","papers":1398,"papersByYear":[2,1,3,10,81,159,311,383,448],"rel":40.9607,"reviewed":true,"share":[0.0305,0.0121,0.0286,0.0841,0.6194,1.0286,1.4635,1.4191,1.2798],"size":1902,"slope":0.002172},{"claims":[147,150,164,181,191,154,290,296,189],"deltaPp":-0.7232,"id":66,"keywords":["stance","news","fake","detection","political"],"label":"Misinformation Stance Detection","labelFull":"Stance Detection for Fake News, Propaganda, and Rumors","papers":822,"papersByYear":[67,72,77,80,85,64,134,138,105],"rel":-0.7068,"reviewed":true,"share":[1.0232,0.8733,0.7335,0.6726,0.6499,0.414,0.6306,0.5113,0.3],"size":1762,"slope":-0.000741},{"claims":[17,22,35,41,52,92,287,467,708],"deltaPp":1.4012,"id":67,"keywords":["reinforcement","agents","preference","reward","learning"],"label":"Reinforcement Learning for LLMs","labelFull":"Reinforcement Learning for Aligning Large Language Models","papers":1313,"papersByYear":[16,13,24,25,30,61,210,358,576],"rel":5.7356,"reviewed":true,"share":[0.2443,0.1577,0.2286,0.2102,0.2294,0.3946,0.9882,1.3265,1.6455],"size":1721,"slope":0.001802},{"claims":[119,109,165,138,148,126,262,246,367],"deltaPp":-0.2501,"id":68,"keywords":["emotion","recognition","affective","multimodal","conversation"],"label":"Multimodal Emotion Recognition","labelFull":"Multimodal Emotion Recognition for Conversational AI","papers":684,"papersByYear":[45,44,61,62,60,51,97,111,153],"rel":-0.3639,"reviewed":true,"share":[0.6872,0.5337,0.5811,0.5212,0.4588,0.3299,0.4564,0.4113,0.4371],"size":1680,"slope":-0.000301},{"claims":[28,61,72,95,123,182,258,426,430],"deltaPp":0.3689,"id":69,"keywords":["anomaly","detection","normal","industrial","video"],"label":"Industrial Visual Anomaly Detection","labelFull":"Visual Anomaly Detection for Industrial Inspection","papers":754,"papersByYear":[17,30,39,43,50,79,104,172,220],"rel":1.421,"reviewed":true,"share":[0.2596,0.3639,0.3715,0.3615,0.3823,0.5111,0.4894,0.6373,0.6285],"size":1675,"slope":0.000447},{"claims":[119,139,144,181,162,241,178,221,274],"deltaPp":-0.4477,"id":70,"keywords":["clustering","k-means","algorithm","approximation","fair"],"label":"Scalable Approximate Clustering","labelFull":"Scalable Approximate Clustering Algorithms","papers":832,"papersByYear":[57,70,65,91,83,107,91,120,148],"rel":-0.5143,"reviewed":true,"share":[0.8705,0.849,0.6192,0.765,0.6347,0.6922,0.4282,0.4446,0.4228],"size":1659,"slope":-0.000576},{"claims":[4,2,6,17,21,50,261,519,773],"deltaPp":1.154,"id":71,"keywords":["hallucination","vision-language","reduces","mitigates","detection"],"label":"Hallucination Detection & Mitigation","labelFull":"Hallucination Detection and Mitigation in LLMs and VLMs","papers":878,"papersByYear":[3,2,6,9,15,28,135,260,420],"rel":25.1965,"reviewed":true,"share":[0.0458,0.0243,0.0572,0.0757,0.1147,0.1811,0.6353,0.9634,1.1998],"size":1653,"slope":0.001449},{"claims":[133,125,178,180,179,188,201,238,228],"deltaPp":-0.5888,"id":72,"keywords":["tracking","trackers","multi-object","visual","state-of-the-art"],"label":"Visual Object Tracking","labelFull":"State-of-the-Art Visual Object Tracking on Benchmarks","papers":765,"papersByYear":[61,50,78,74,76,89,98,119,120],"rel":-0.632,"reviewed":true,"share":[0.9316,0.6064,0.7431,0.6221,0.5811,0.5758,0.4612,0.4409,0.3428],"size":1650,"slope":-0.000577},{"claims":[70,96,110,146,145,197,207,295,367],"deltaPp":0.0247,"id":73,"keywords":["generative","modeling","latent","deep","distribution"],"label":"Plug-and-Play Generative Modeling","labelFull":"Plug-and-Play Adaptation for Efficient Generative Models","papers":1272,"papersByYear":[56,69,81,118,111,141,159,229,308],"rel":0.0289,"reviewed":true,"share":[0.8552,0.8369,0.7716,0.992,0.8488,0.9121,0.7482,0.8485,0.8799],"size":1633,"slope":1e-06},{"claims":[124,144,119,135,79,149,195,258,369],"deltaPp":-0.2777,"id":74,"keywords":["low-rank","matrix","matrices","factorization","sensing"],"label":"Low-Rank Adaptation","labelFull":"Low-Rank Adaptation for Efficient Model Fine-Tuning","papers":1014,"papersByYear":[70,73,79,74,50,90,122,179,277],"rel":-0.2598,"reviewed":true,"share":[1.069,0.8854,0.7526,0.6221,0.3823,0.5822,0.5741,0.6632,0.7913],"size":1572,"slope":-0.000362},{"claims":[71,128,135,207,176,196,203,218,220],"deltaPp":-0.1363,"id":75,"keywords":["style","stylization","transfer","content","font"],"label":"Arbitrary Style Transfer","labelFull":"Arbitrary Image Style Transfer and Stylization","papers":806,"papersByYear":[31,57,69,101,102,97,121,110,118],"rel":-0.2879,"reviewed":true,"share":[0.4734,0.6913,0.6573,0.8491,0.7799,0.6275,0.5694,0.4076,0.3371],"size":1554,"slope":-0.000299},{"claims":[280,244,250,198,194,116,136,61,36],"deltaPp":-2.0657,"id":76,"keywords":["parsing","parser","dependency","constituency","semantic"],"label":"Dependency Parsing","labelFull":"Neural Dependency and Semantic Dependency Parsing","papers":748,"papersByYear":[139,111,113,103,101,57,67,37,20],"rel":-0.9731,"reviewed":true,"share":[2.1228,1.3463,1.0765,0.8659,0.7723,0.3687,0.3153,0.1371,0.0571],"size":1515,"slope":-0.002318},{"claims":[58,100,127,147,174,155,271,263,209],"deltaPp":-0.1911,"id":77,"keywords":["text","generation","natural","detection","evaluation"],"label":"Neural Text Generation","labelFull":"Neural Text Generation: Methods, Benchmarks, and Evaluation","papers":1042,"papersByYear":[43,71,81,95,118,107,188,176,163],"rel":-0.291,"reviewed":true,"share":[0.6567,0.8611,0.7716,0.7987,0.9023,0.6922,0.8847,0.6521,0.4656],"size":1504,"slope":-0.000212},{"claims":[20,43,79,121,124,186,244,249,436],"deltaPp":0.7793,"id":78,"keywords":["uncertainty","calibration","out-of-distribution","epistemic","quantification"],"label":"Uncertainty Quantification","labelFull":"Uncertainty Quantification and Calibration in Machine Learning","papers":1181,"papersByYear":[15,27,59,93,107,150,192,185,353],"rel":3.4016,"reviewed":true,"share":[0.2291,0.3275,0.5621,0.7818,0.8182,0.9704,0.9035,0.6855,1.0084],"size":1502,"slope":0.000844},{"claims":[153,175,198,202,138,91,144,192,196],"deltaPp":-0.7022,"id":79,"keywords":["person","re-identification","identity","visible-infrared","search"],"label":"Person Re-identification","labelFull":"Person Re-identification (Image, Video, and Cross-modality)","papers":736,"papersByYear":[66,86,93,93,67,49,78,97,107],"rel":-0.6967,"reviewed":true,"share":[1.0079,1.0431,0.886,0.7818,0.5123,0.317,0.367,0.3594,0.3057],"size":1489,"slope":-0.00106},{"claims":[111,130,162,199,185,203,213,171,105],"deltaPp":-0.6915,"id":80,"keywords":["event","extraction","argument","biomedical","relation"],"label":"Event Extraction","labelFull":"Event Extraction and Event Relation Extraction","papers":748,"papersByYear":[58,67,85,93,88,92,104,93,68],"rel":-0.7807,"reviewed":true,"share":[0.8858,0.8126,0.8098,0.7818,0.6729,0.5952,0.4894,0.3446,0.1943],"size":1479,"slope":-0.000833},{"claims":[77,110,131,161,163,197,169,201,267],"deltaPp":-0.3614,"id":81,"keywords":["complexity","algorithm","bounds","lower","log"],"label":"Algorithmic Complexity Analysis","labelFull":"Algorithmic Complexity Analysis and Theoretical Guarantees","papers":1248,"papersByYear":[65,90,113,137,142,169,147,164,221],"rel":-0.3641,"reviewed":true,"share":[0.9927,1.0916,1.0765,1.1517,1.0858,1.0933,0.6917,0.6077,0.6313],"size":1476,"slope":-0.000621},{"claims":[2,6,20,29,54,84,298,401,545],"deltaPp":0.5018,"id":82,"keywords":["watermark","steganography","robustness","embeds","image"],"label":"Robust AI Watermarking","labelFull":"Robust Watermarking for AI-Generated Text, Images, and Video","papers":473,"papersByYear":[1,3,7,10,18,29,95,129,181],"rel":32.7974,"reviewed":true,"share":[0.0153,0.0364,0.0667,0.0841,0.1376,0.1876,0.447,0.478,0.5171],"size":1439,"slope":0.000699},{"claims":[48,45,90,78,94,84,160,269,533],"deltaPp":0.5628,"id":83,"keywords":["clinical","medical","disease","patient","diagnosis"],"label":"Clinical Prediction Models","labelFull":"Machine Learning Models for Clinical Prediction and Diagnosis","papers":855,"papersByYear":[26,27,47,48,55,62,93,161,336],"rel":1.4173,"reviewed":true,"share":[0.3971,0.3275,0.4477,0.4035,0.4206,0.4011,0.4376,0.5965,0.9599],"size":1401,"slope":0.000506},{"claims":[50,73,86,108,143,197,193,222,325],"deltaPp":0.155,"id":84,"keywords":["equivariant","manifold","group","networks","space"],"label":"Equivariant Manifold Networks","labelFull":"Group Equivariant Neural Networks on Manifolds","papers":921,"papersByYear":[34,45,53,73,88,121,129,142,236],"rel":0.2985,"reviewed":true,"share":[0.5192,0.5458,0.5049,0.6137,0.6729,0.7828,0.607,0.5261,0.6742],"size":1397,"slope":0.000156},{"claims":[120,163,165,112,116,115,254,183,125],"deltaPp":-0.6533,"id":85,"keywords":["discourse","argument","essay","mining","scoring"],"label":"Argument Mining","labelFull":"Argument Mining: Extraction, Relation Prediction, Scoring","papers":597,"papersByYear":[54,71,76,61,52,50,105,68,60],"rel":-0.7922,"reviewed":true,"share":[0.8247,0.8611,0.724,0.5128,0.3976,0.3235,0.4941,0.252,0.1714],"size":1353,"slope":-0.000848},{"claims":[54,60,71,80,83,141,175,317,355],"deltaPp":0.0846,"id":86,"keywords":["multi-view","clustering","cross-view","incomplete","consistency"],"label":"Multi-View Clustering","labelFull":"Deep Multi-View Clustering with Incomplete and Noisy Views","papers":701,"papersByYear":[30,34,42,42,49,73,89,152,190],"rel":0.1846,"reviewed":true,"share":[0.4582,0.4124,0.4001,0.3531,0.3747,0.4722,0.4188,0.5632,0.5428],"size":1336,"slope":0.000158},{"claims":[79,92,138,151,161,201,184,167,160],"deltaPp":-0.5564,"id":87,"keywords":["ensemble","loss","boosting","bounds","surrogate"],"label":"Surrogate Loss Theory","labelFull":"Theoretical Guarantees and Generalization Bounds for Surrogate Losses","papers":935,"papersByYear":[60,68,96,104,114,134,114,119,126],"rel":-0.6072,"reviewed":true,"share":[0.9163,0.8247,0.9145,0.8743,0.8717,0.8669,0.5364,0.4409,0.3599],"size":1333,"slope":-0.00069},{"claims":[70,77,95,88,127,207,214,221,198],"deltaPp":-0.0675,"id":88,"keywords":["hand","gesture","pose","3d","contact"],"label":"3D Hand-Object Pose Estimation","labelFull":"3D Hand Pose and Hand-Object Interaction Estimation","papers":514,"papersByYear":[22,26,33,35,48,75,92,89,94],"rel":-0.2009,"reviewed":true,"share":[0.336,0.3153,0.3144,0.2942,0.367,0.4852,0.4329,0.3298,0.2685],"size":1297,"slope":3.4e-05},{"claims":[43,50,63,119,113,148,193,259,289],"deltaPp":0.0597,"id":89,"keywords":["retrieval","image","composed","cross-modal","zero-shot"],"label":"Cross-Modal Retrieval","labelFull":"Cross-Modal Retrieval for Images, Videos, and Text","papers":897,"papersByYear":[35,34,39,82,75,112,135,177,208],"rel":0.1117,"reviewed":true,"share":[0.5345,0.4124,0.3715,0.6894,0.5735,0.7245,0.6353,0.6558,0.5942],"size":1277,"slope":0.000255},{"claims":[50,61,117,142,158,158,173,178,238],"deltaPp":-0.0208,"id":90,"keywords":["fairness","group","constraints","guarantees","empirical"],"label":"Fairness-Accuracy Trade-Offs","labelFull":"Fairness-Accuracy Trade-Offs in Machine Learning","papers":651,"papersByYear":[24,30,49,74,77,78,102,96,121],"rel":-0.0568,"reviewed":true,"share":[0.3665,0.3639,0.4668,0.6221,0.5888,0.5046,0.48,0.3557,0.3457],"size":1275,"slope":-3.3e-05},{"claims":[17,34,37,30,55,94,144,287,555],"deltaPp":0.5082,"id":91,"keywords":["eeg","brain","fmri","signals","decoding"],"label":"EEG Foundation Models","labelFull":"EEG Foundation Models: Pretraining and Benchmarking","papers":534,"papersByYear":[9,17,17,17,28,47,57,116,226],"rel":3.6987,"reviewed":true,"share":[0.1374,0.2062,0.162,0.1429,0.2141,0.304,0.2682,0.4298,0.6456],"size":1253,"slope":0.000513},{"claims":[55,98,135,125,141,141,135,160,246],"deltaPp":-0.1406,"id":92,"keywords":["estimator","variance","error","asymptotic","derived"],"label":"Estimation Theory","labelFull":"Estimation Theory with Finite‑Sample Guarantees","papers":917,"papersByYear":[44,66,94,92,107,101,103,124,186],"rel":-0.2092,"reviewed":true,"share":[0.672,0.8005,0.8955,0.7734,0.8182,0.6534,0.4847,0.4594,0.5314],"size":1236,"slope":-0.000421},{"claims":[25,29,39,35,39,59,146,318,544],"deltaPp":0.8259,"id":93,"keywords":["human","judgments","evaluation","judges","llm"],"label":"LLM-as-a-Judge","labelFull":"LLM-as-a-Judge Evaluation Aligned with Human Judgments","papers":950,"papersByYear":[20,28,35,33,39,50,122,227,396],"rel":2.7043,"reviewed":true,"share":[0.3054,0.3396,0.3334,0.2774,0.2982,0.3235,0.5741,0.8411,1.1313],"size":1234,"slope":0.000889},{"claims":[21,24,75,95,96,114,141,251,403],"deltaPp":0.7261,"id":94,"keywords":["trajectory","prediction","travel","future","state-of-the-art"],"label":"Trajectory Forecasting","labelFull":"Trajectory Forecasting for Autonomous Driving","papers":845,"papersByYear":[14,17,45,56,58,75,100,151,329],"rel":3.3962,"reviewed":true,"share":[0.2138,0.2062,0.4287,0.4708,0.4435,0.4852,0.4706,0.5595,0.9399],"size":1220,"slope":0.000677},{"claims":[34,49,88,128,145,165,168,174,257],"deltaPp":0.2811,"id":95,"keywords":["student","teacher","knowledge","network","trained"],"label":"Knowledge Distillation","labelFull":"Teacher-Student Knowledge Distillation for Model Compression","papers":867,"papersByYear":[19,26,59,84,103,123,119,134,200],"rel":0.9686,"reviewed":true,"share":[0.2902,0.3153,0.5621,0.7062,0.7876,0.7957,0.56,0.4965,0.5713],"size":1208,"slope":0.000292},{"claims":[185,216,201,221,148,91,59,22,21],"deltaPp":-1.2705,"id":96,"keywords":["gan","discriminator","generator","wasserstein","conditional"],"label":"Stable GAN Training","labelFull":"Techniques for Stable and Efficient GAN Training","papers":633,"papersByYear":[86,112,108,118,92,55,31,16,15],"rel":-0.9673,"reviewed":true,"share":[1.3134,1.3584,1.0289,0.992,0.7035,0.3558,0.1459,0.0593,0.0429],"size":1164,"slope":-0.001897},{"claims":[12,17,27,21,16,29,169,405,463],"deltaPp":0.7653,"id":97,"keywords":["preference","optimization","direct","alignment","pairs"],"label":"Direct Preference Optimization","labelFull":"Direct Preference Optimization for LLM Alignment","papers":799,"papersByYear":[9,13,22,12,12,21,106,288,316],"rel":5.5699,"reviewed":true,"share":[0.1374,0.1577,0.2096,0.1009,0.0918,0.1359,0.4988,1.0671,0.9027],"size":1159,"slope":0.001067}],"totals":{"claims":[29959,38280,49207,56294,62223,74287,102084,133164,180022],"papers":[6548,8245,10497,11895,13078,15458,21251,26989,35005]},"years":[2018,2019,2020,2021,2022,2023,2024,2025,2026]}