Top 10 Arxiv Papers Today in Machine Learning


2.388 Mikeys
#1. Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model
Julian Schrittwieser, Ioannis Antonoglou, Thomas Hubert, Karen Simonyan, Laurent Sifre, Simon Schmitt, Arthur Guez, Edward Lockhart, Demis Hassabis, Thore Graepel, Timothy Lillicrap, David Silver
Constructing agents with planning capabilities has long been one of the main challenges in the pursuit of artificial intelligence. Tree-based planning methods have enjoyed huge success in challenging domains, such as chess and Go, where a perfect simulator is available. However, in real-world problems the dynamics governing the environment are often complex and unknown. In this work we present the MuZero algorithm which, by combining a tree-based search with a learned model, achieves superhuman performance in a range of challenging and visually complex domains, without any knowledge of their underlying dynamics. MuZero learns a model that, when applied iteratively, predicts the quantities most directly relevant to planning: the reward, the action-selection policy, and the value function. When evaluated on 57 different Atari games - the canonical video game environment for testing AI techniques, in which model-based planning approaches have historically struggled - our new algorithm achieved a new state of the art. When evaluated...
more | pdf | html
Figures
Tweets
mathena: Holy Cow this is a good step towards AGI. https://t.co/LPVSTxu08J
HNTweets: MuZero beats AlphaZero, with less training and no explicit rules: fully general: https://t.co/S28mwg5ZYk Comments: https://t.co/LbP3sARyUS
angsuman: MuZero beats AlphaZero, with less training and no explicit rules: fully general https://t.co/sjhWZdrNTP
polynoamial: "Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model" -- Really exciting new work from the AlphaZero group at DeepMind! https://t.co/IJAwdsAc6q
PhilippBayer: This looks very exciting! MuZero, matches the best (specialised?) Shogi and Chess algorithms, best in Go, but also best in 57 Atari games! https://t.co/DZ15zt6c7h Introduces a mix of planning algorithms and reinforcement learning (cc @PerthMLGroup)
jaguring1: 二日前にディープマインドから新しい論文。囲碁でプロに対し圧倒的な強さを見せつけたアルファ碁、アルファ碁ゼロ、アルファゼロの開発にも関わった、あのシュリットヴィーザーらの研究。名前はMuZero Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/I2w79hgXRs https://t.co/YES1mrzfRf
IntuitMachine: DeepMind is so ahead of the curve. I was dreaming last night on a novel way to formulate an RL solution. Only find out this morning, that DeepMind implemented my dream and has a paper out! https://t.co/xlpgPAXhQC .
BrundageBot: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model. Schrittwieser, Antonoglou, Hubert, Simonyan, Sifre, Schmitt, Guez, Lockhart, Hassabis, Graepel, Lillicrap, and Silver https://t.co/gWUdaaVgg5
rinatie_ceo: [1911.08265] Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/06Ij9ZXx69
jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければいけない情報の量を大幅に削減しててとても美しい。 https://t.co/tTSxpQIgZY
federicolois: A new year, a new AlphaZero paper. Muzero now in model-free flavour without losing capability. #reinforcementlearning https://t.co/hOdxjBbNsH
shtoons: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model. Crazy stuff. https://t.co/AmyxyVF4wT
rosinality: https://t.co/ymLJ2gITxL 오 이젠 알파고가 아타리도 하네. 모델 기반 RL로의 확장이라 흥미로움.
jreuben1: "MuZero: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model", Schrittwieser et al 2019 https://t.co/54bjhou2a5 [tree search over learned latent-dynamics model reaches AlphaZero level; plus beating R2D2 & SimPLe ALE SOTAs]
hn_frontpage: MuZero beats AlphaZero, with less training and no explicit rules: fully general L: https://t.co/ar7Avd2ahU C: https://t.co/DXtEyW7h30
jafbm57: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model #AI https://t.co/IVjoxCOmZ1
ankesh_anand: Exciting new paper from @DeepMindAI : Planning with **learned** models can scale to complex visual domains like Atari. TL;DR: MCTS + Q-learning with models that only predict rewards/Q-values leads to new SOTA on Atari games and matches AlphaZero on Go. https://t.co/5fNWk3z7nD https://t.co/d961KoFiiy
KloudStrife: MuZero - MCTS on Atari, chess-Go, SotA, !! no knowledge of game rules or environment dynamics required !! Key paper. https://t.co/EUA4Ey8w2s
hackernewsj: 学習モデルを使用した計画によるアタリ、囲Go、チェス、将giの習得 https://t.co/sYDQ6MFGj2
therealjpittman: [R] [1911.08265] Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/llTAMvEb2e #MachineLearning
therealjpittman: [1911.08265] Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model | Arxiv https://t.co/llTAMvEb2e #AI #ArtificialIntelligence
cute_na_piglets: ついに将棋AIから詰みの概念が完全に無くなった! 先読みらしいことをやっているがもはや局面も手も抽象的な何かに過ぎず、思考内容を解釈することは不可能。 https://t.co/65r05NZoLB
cute_na_piglets: 訂正 抽象的な局面空間の中で先読みを行なっている。ルールを直接使わないだけで手はわかるので、おそらく思考内容を解釈することも可能なはずです。 https://t.co/65r05NZoLB
jonathanrraiman: Incredible work on Model-based RL that finally outperforms other approaches on both continuous/visual games (Atari) and board games (go, chess, shogu) from @DeepMindAI @Mononofu, Antonoglou, Hubert et al. So many problems can be cast this way, congrats! https://t.co/HxRxgvekTA
k_matsuzaki: 囲碁だと60万ステップでAlphaZeroに勝てるらしいのでUEC杯までに誰か実装してくるに違いない。100並列くらいで多分間に合う。 https://t.co/BtidWq3E5R
reddit_ml: [R] [1911.08265] Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/dAo1VRjYHo
a_tschantz: State-of-the-art in Atari with a model-based approach - https://t.co/5zqf6MXUPS
Bordeaux_007: 【メモ】明日読む Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/gL3oRiRE4Y
hacker_news_hir: MuZero beats AlphaZero, with less training and no explicit rules: fully general : https://t.co/aUWMqfrqXb Comments: https://t.co/EZdpukaTiF
mohapsat: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/zaQmPCKTzq #AI #news #tech
autonomyEV: @strangecosmos @Cruise might find this interesting Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/h4GkzggFQG
HackerNewsPosts: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model Link: https://t.co/FQ7G1JWmrj Cmts: https://t.co/zTMO7O3f0m
arankomatsuzaki: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/iiVt0FInKj MuZero combines a tree-based search with a learned model and achieves superhuman performance, without knowledge of dynamics, at various games, including Go and Atari (sota).
zarzuelazen: Paper: "Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model" https://t.co/mZeXdaxnEM
betterhn50: 52 – Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/yrOIEUXq8y
octonion: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model - https://t.co/XRzZkqAta7
y_ich2: またDeepMind。 ルールを教えなくてもAlphaZeroと同等の強さになったと。 [1911.08265] Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/iTy96MSt5w
StephenPiment: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/BI9LR8NtVd
StatsPapers: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model. https://t.co/kuLNUldJLn
william_woof: Okay, we all knew that model-based approaches are going to beat model-free approaches, but Deepmind's new paper suggests that (learned) model-based approaches can even beat using the ground truth model directly: https://t.co/6DWPZFrWLk
SMBrocklehurst: Great to see @DeepMindAI keeps moving forward with reinforcement learning. This is a nice piece of work - "Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model". https://t.co/yZVf7Z4JX8
an_interstice: New paper by DeepMind looks very interesting. Using planning with a learned model, they build SOTA agents for Atari, and achieve superhuman performance at Go, chess and shogi: https://t.co/RPBDWfo6be
mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
hackernews100: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/0qntbv60xj
hackernewsfeed: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/icnpk4XZlF
PolyaxonAI: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/vLdMe6uwsq
hackernewsrobot: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/7oRTk30IsK
autonomyEV: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/h4GkzggFQG
butlersean: https://t.co/hq54ewhjze
KaiLashArul: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
rinatie_ceo: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
oubeika11: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
KazuSamejima: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
ararabo: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
kmo2: RT @cute_na_piglets: ついに将棋AIから詰みの概念が完全に無くなった! 先読みらしいことをやっているがもはや局面も手も抽象的な何かに過ぎず、思考内容を解釈することは不可能。 https://t.co/65r05NZoLB
nishinojunji: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
tak_yamm: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
morioka: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
morioka: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
atsushi_craft: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
turkeyfiend: RT @rosinality: https://t.co/ymLJ2gITxL 오 이젠 알파고가 아타리도 하네. 모델 기반 RL로의 확장이라 흥미로움.
cute_na_piglets: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
cute_na_piglets: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
tensword: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
xpasky: RT @arankomatsuzaki: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/iiVt0FInKj MuZero combines a tree-…
ETCShogi: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
qwazpia: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
k_matsuzaki: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
63556poiuytrewq: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
am_nimitz3: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
dannyehb: RT @arankomatsuzaki: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/iiVt0FInKj MuZero combines a tree-…
ihme_vaeltaa: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
MarkTan57229491: RT @arankomatsuzaki: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/iiVt0FInKj MuZero combines a tree-…
veydpz_public: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
zot_msh: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
pranjaltandon2: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
kz_lil_fox: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
istaaaaaaaaa: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
THsama2: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
makiedan: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
MoTaylor95: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
kli_nlpr: RT @mooopan: Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model https://t.co/wvd4kKaSpc 😲
miki_iwa: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
AROMABL4CK: RT @cute_na_piglets: ついに将棋AIから詰みの概念が完全に無くなった! 先読みらしいことをやっているがもはや局面も手も抽象的な何かに過ぎず、思考内容を解釈することは不可能。 https://t.co/65r05NZoLB
Mt_El_Sheep: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
vvmatorin: RT @mathena: Holy Cow this is a good step towards AGI. https://t.co/LPVSTxu08J
harunahaju: RT @jinbeizame007: MuZeroの論文(https://t.co/tZ9HWB8Tb7)面白いな。状態から行動, 状態価値, 報酬までEnd-to-Endで学習しつつ、そのモデルの隠れ状態のダイナミクスモデルを学習することで、ダイナミクスモデルが予測しなければ…
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 12
Total Words: 10432
Unqiue Words: 3522

2.099 Mikeys
#2. Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking
Eric Crawford, Joelle Pineau
The ability to detect and track objects in the visual world is a crucial skill for any intelligent agent, as it is a necessary precursor to any object-level reasoning process. Moreover, it is important that agents learn to track objects without supervision (i.e. without access to annotated training videos) since this will allow agents to begin operating in new environments with minimal human assistance. The task of learning to discover and track objects in videos, which we call \textit{unsupervised object tracking}, has grown in prominence in recent years; however, most architectures that address it still struggle to deal with large scenes containing many objects. In the current work, we propose an architecture that scales well to the large-scene, many-object setting by employing spatially invariant computations (convolutions and spatial attention) and representations (a spatially local object specification scheme). In a series of experiments, we demonstrate a number of attractive features of our architecture; most notably, that...
more | pdf | html
Figures
None.
Tweets
BrundageBot: Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking. Eric Crawford and Joelle Pineau https://t.co/JIlxKej9dt
arxivml: "Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking", Eric Crawford, Joelle Pineau https://t.co/e06BiCgprc
arxiv_cs_LG: Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking. Eric Crawford and Joelle Pineau https://t.co/EfgEvRfAgj
StatsPapers: Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking. https://t.co/MOwyqJ8vXq
arxiv_cscv: Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking https://t.co/GMYn4Sri7O
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 0
Unqiue Words: 0

2.099 Mikeys
#3. 3D-Rotation-Equivariant Quaternion Neural Networks
Binbin Zhang, Wen Shen, Shikun Huang, Zhihua Wei, Quanshi Zhang
This paper proposes a set of rules to revise various neural networks for 3D point cloud processing to rotation-equivariant quaternion neural networks (REQNNs). We find that when a neural network uses quaternion features under certain conditions, the network feature naturally has the rotation-equivariance property. Rotation equivariance means that applying a specific rotation transformation to the input point cloud is equivalent to applying the same rotation transformation to all intermediate-layer quaternion features. Besides, the REQNN also ensures that the intermediate-layer features are invariant to the permutation of input points. Compared with the original neural network, the REQNN exhibits higher rotation robustness.
more | pdf | html
Figures
None.
Tweets
BrundageBot: 3D-Rotation-Equivariant Quaternion Neural Networks. Binbin Zhang, Wen Shen, Shikun Huang, Zhihua Wei, and Quanshi Zhang https://t.co/VaUMbKQ6dI
arxivml: "3D-Rotation-Equivariant Quaternion Neural Networks", Binbin Zhang, Wen Shen, Shikun Huang, Zhihua Wei, Quanshi Zha… https://t.co/HtwuB36NQd
arxiv_cs_LG: 3D-Rotation-Equivariant Quaternion Neural Networks. Binbin Zhang, Wen Shen, Shikun Huang, Zhihua Wei, and Quanshi Zhang https://t.co/vMiyMzLp8e
StatsPapers: 3D-Rotation-Equivariant Quaternion Neural Networks. https://t.co/AoBi3N1r6h
arxiv_cscv: 3D-Rotation-Equivariant Quaternion Neural Networks https://t.co/7GBUukfGgJ
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 5
Total Words: 0
Unqiue Words: 0

2.099 Mikeys
#4. Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning
Junjie Wang, Xiangfeng Wang, Bo Jin, Junchi Yan, Wenjie Zhang, Hongyuan Zha
Generalized zero-shot learning (GZSL) tackles the problem of learning to classify instances involving both seen classes and unseen ones. The key issue is how to effectively transfer the model learned from seen classes to unseen classes. Existing works in GZSL usually assume that some prior information about unseen classes are available. However, such an assumption is unrealistic when new unseen classes appear dynamically. To this end, we propose a novel heterogeneous graph-based knowledge transfer method (HGKT) for GZSL, agnostic to unseen classes and instances, by leveraging graph neural network. Specifically, a structured heterogeneous graph is constructed with high-level representative nodes for seen classes, which are chosen through Wasserstein barycenter in order to simultaneously capture inter-class and intra-class relationship. The aggregation and embedding functions can be learned through graph neural network, which can be used to compute the embeddings of unseen classes by transferring the knowledge from their neighbors....
more | pdf | html
Figures
Tweets
BrundageBot: Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning. Junjie Wang, Xiangfeng Wang, Bo Jin, Junchi Yan, Wenjie Zhang, and Hongyuan Zha https://t.co/LdApZW0h6j
arxivml: "Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning", Junjie Wang, Xiangfeng Wang, Bo … https://t.co/uHBsC0NYyb
arxiv_cs_LG: Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning. Junjie Wang, Xiangfeng Wang, Bo Jin, Junchi Yan, Wenjie Zhang, and Hongyuan Zha https://t.co/j6R7eTWaSD
StatsPapers: Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning. https://t.co/z5Bq9TfhXa
arxiv_cscv: Heterogeneous Graph-based Knowledge Transfer for Generalized Zero-shot Learning https://t.co/TVsHTV6wJ6
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 6
Total Words: 6377
Unqiue Words: 2270

2.098 Mikeys
#5. Exponential Family Graph Embeddings
Abdulkadir Çelikkanat, Fragkiskos D. Malliaros
Representing networks in a low dimensional latent space is a crucial task with many interesting applications in graph learning problems, such as link prediction and node classification. A widely applied network representation learning paradigm is based on the combination of random walks for sampling context nodes and the traditional \textit{Skip-Gram} model to capture center-context node relationships. In this paper, we emphasize on exponential family distributions to capture rich interaction patterns between nodes in random walk sequences. We introduce the generic \textit{exponential family graph embedding} model, that generalizes random walk-based network representation learning techniques to exponential family conditional distributions. We study three particular instances of this model, analyzing their properties and showing their relationship to existing unsupervised learning models. Our experimental evaluation on real-world datasets demonstrates that the proposed techniques outperform well-known baseline methods in two...
more | pdf | html
Figures
None.
Tweets
BrundageBot: Exponential Family Graph Embeddings. Abdulkadir Çelikkanat and Fragkiskos D. Malliaros https://t.co/wqB6kpG99C
arxivml: "Exponential Family Graph Embeddings", Abdulkadir Çelikkanat, Fragkiskos D. Malliaros https://t.co/sByXJyyweR
arxiv_cs_LG: Exponential Family Graph Embeddings. Abdulkadir Çelikkanat and Fragkiskos D. Malliaros https://t.co/O9yMmAvI7S
StatsPapers: Exponential Family Graph Embeddings. https://t.co/jjnPvncdQh
RexDouglass: RT @StatsPapers: Exponential Family Graph Embeddings. https://t.co/jjnPvncdQh
mizvladimir: RT @StatsPapers: Exponential Family Graph Embeddings. https://t.co/jjnPvncdQh
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 9018
Unqiue Words: 2469

2.094 Mikeys
#6. Outside the Box: Abstraction-Based Monitoring of Neural Networks
Thomas A. Henzinger, Anna Lukina, Christian Schilling
Neural networks have demonstrated unmatched performance in a range of classification tasks. Despite numerous efforts of the research community, novelty detection remains one of the significant limitations of neural networks. The ability to identify previously unseen inputs as novel is crucial for our understanding of the decisions made by neural networks. At runtime, inputs not falling into any of the categories learned during training cannot be classified correctly by the neural network. Existing approaches treat the neural network as a black box and try to detect novel inputs based on the confidence of the output predictions. However, neural networks are not trained to reduce their confidence for novel inputs, which limits the effectiveness of these approaches. We propose a framework to monitor a neural network by observing the hidden layers. We employ a common abstraction from program analysis - boxes - to identify novel behaviors in the monitored layers, i.e., inputs that cause behaviors outside the box. For each neuron, the...
more | pdf | html
Figures
None.
Tweets
BrundageBot: Outside the Box: Abstraction-Based Monitoring of Neural Networks. Thomas A. Henzinger, Anna Lukina, and Christian Schilling https://t.co/EWebNqQvi1
arxivml: "Outside the Box: Abstraction-Based Monitoring of Neural Networks", Thomas A. Henzinger, Anna Lukina, Christian Sch… https://t.co/mAtpmZ7Ts4
arxiv_cs_LG: Outside the Box: Abstraction-Based Monitoring of Neural Networks. Thomas A. Henzinger, Anna Lukina, and Christian Schilling https://t.co/40F8GiHf1k
SciFi: Outside the Box: Abstraction-Based Monitoring of Neural Networks. https://t.co/t8fH9Jere5
arxiv_cslo: Outside the Box: Abstraction-Based Monitoring of Neural Networks https://t.co/F78A0v1kIt
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 3
Total Words: 0
Unqiue Words: 0

2.079 Mikeys
#7. Local AdaAlter: Communication-Efficient Stochastic Gradient Descent with Adaptive Learning Rates
Cong Xie, Oluwasanmi Koyejo, Indranil Gupta, Haibin Lin
Recent years have witnessed the growth of large-scale distributed machine learning algorithms -- specifically designed to accelerate model training by distributing computation across multiple machines. When scaling distributed training in this way, the communication overhead is often the bottleneck. In this paper, we study the local distributed Stochastic Gradient Descent~(SGD) algorithm, which reduces the communication overhead by decreasing the frequency of synchronization. While SGD with adaptive learning rates is a widely adopted strategy for training neural networks, it remains unknown how to implement adaptive learning rates in local SGD. To this end, we propose a novel SGD variant with reduced communication and adaptive learning rates, with provable convergence. Empirical results show that the proposed algorithm has fast convergence and efficiently reduces the communication overhead.
more | pdf | html
Figures
None.
Tweets
BrundageBot: Local AdaAlter: Communication-Efficient Stochastic Gradient Descent with Adaptive Learning Rates. Cong Xie, Oluwasanmi Koyejo, Indranil Gupta, and Haibin Lin https://t.co/g2KCILbY0y
arxivml: "Local AdaAlter: Communication-Efficient Stochastic Gradient Descent with Adaptive Learning Rates", Cong Xie, Oluwa… https://t.co/e5TZHW8nZP
arxiv_cs_LG: Local AdaAlter: Communication-Efficient Stochastic Gradient Descent with Adaptive Learning Rates. Cong Xie, Oluwasanmi Koyejo, Indranil Gupta, and Haibin Lin https://t.co/IY5RWqIWUL
StatsPapers: Local AdaAlter: Communication-Efficient Stochastic Gradient Descent with Adaptive Learning Rates. https://t.co/uLe51EnxJf
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 4
Total Words: 0
Unqiue Words: 0

2.079 Mikeys
#8. Adaptive Wind Driven Optimization Trained Artificial Neural Networks
Zikri Bayraktar
This paper presents the application of a newly developed nature-inspired metaheuristic optimization method, namely the Adaptive Wind Driven Optimization (AWDO), to the training of feedforward artificial neural networks (NN) and presents a discussion into the future research of AWDO implementation in Deep Learning (DL). Application example of digit classification with MNIST dataset reveals interesting behavior of the derivative-free AWDO method compared to steepest descent method where results and future work on the implementation of AWDO in deep neural networks are discussed.
more | pdf | html
Figures
None.
Tweets
BrundageBot: Adaptive Wind Driven Optimization Trained Artificial Neural Networks. Zikri Bayraktar https://t.co/j4vO4QERSk
arxivml: "Adaptive Wind Driven Optimization Trained Artificial Neural Networks", Zikri Bayraktar https://t.co/X4oBSxL8gY
arxiv_cs_LG: Adaptive Wind Driven Optimization Trained Artificial Neural Networks. Zikri Bayraktar https://t.co/BZ354Nb9hw
StatsPapers: Adaptive Wind Driven Optimization Trained Artificial Neural Networks. https://t.co/DbJzMz1gEZ
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 1
Total Words: 0
Unqiue Words: 0

2.074 Mikeys
#9. Shapelets for earthquake detection
Monica Arul, Ahsan Kareem
This paper introduces EQShapelets (EarthQuake Shapelets) a time-series shape-based approach embedded in machine learning to autonomously detect earthquakes. It promises to overcome the challenges in the field of seismology related to automated detection and cataloging of earthquakes. EQShapelets are amplitude and phase-independent, i.e., their detection sensitivity is irrespective of the magnitude of the earthquake and the time of occurrence. They are also robust to noise and other spurious signals. The detection capability of EQShapelets is tested on one week of continuous seismic data provided by the Northern California Seismic Network (NCSN) obtained from a station in central California near the Calaveras Fault. EQShapelets combined with a Random Forest classifier, detected all of the cataloged earthquakes and 281 uncataloged events with lower false detection rate thus offering a better performance than autocorrelation and FAST algorithms. The primary advantage of EQShapelets over competing methods is the interpretability and...
more | pdf | html
Figures
Tweets
arxivml: "Shapelets for earthquake detection", Monica Arul, Ahsan Kareem https://t.co/IDjV9r6gRJ
arxiv_cs_LG: Shapelets for earthquake detection. Monica Arul and Ahsan Kareem https://t.co/qFIE1wkbiH
jahalat: [1911.09086] Shapelets for earthquake detection https://t.co/r48CNrxYSV
StatsPapers: Shapelets for earthquake detection. https://t.co/RvwA3joeRJ
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 2
Total Words: 5959
Unqiue Words: 1875

2.073 Mikeys
#10. Towards a Unified Evaluation of Explanation Methods without Ground Truth
Hao Zhang, Jiayi Chen, Haotian Xue, Quanshi Zhang
This paper proposes a set of criteria to evaluate the objectiveness of explanation methods of neural networks, which is crucial for the development of explainable AI, but it also presents significant challenges. The core challenge is that people usually cannot obtain ground-truth explanations of the neural network. To this end, we design four metrics to evaluate explanation results without ground-truth explanations. Our metrics can be broadly applied to nine benchmark methods of interpreting neural networks, which provides new insights of explanation methods.
more | pdf | html
Figures
None.
Tweets
arxivml: "Towards a Unified Evaluation of Explanation Methods without Ground Truth", Hao Zhang, Jiayi Chen, Haotian Xue, Qua… https://t.co/ZRZ5i3qZBL
arxiv_cs_LG: Towards a Unified Evaluation of Explanation Methods without Ground Truth. Hao Zhang, Jiayi Chen, Haotian Xue, and Quanshi Zhang https://t.co/Flg7D7VgSU
SciFi: Towards a Unified Evaluation of Explanation Methods without Ground Truth. https://t.co/UoizF3SoKH
arxiv_cscv: Towards a Unified Evaluation of Explanation Methods without Ground Truth https://t.co/hCOMiT8xnl
Github
None.
Youtube
None.
Other stats
Sample Sizes : None.
Authors: 4
Total Words: 0
Unqiue Words: 0

About

Assert is a website where the best academic papers on arXiv (computer science, math, physics), bioRxiv (biology), BITSS (reproducibility), EarthArXiv (earth science), engrXiv (engineering), LawArXiv (law), PsyArXiv (psychology), SocArXiv (social science), and SportRxiv (sport research) bubble to the top each day.

Papers are scored (in real-time) based on how verifiable they are (as determined by their Github repos) and how interesting they are (based on Twitter).

To see top papers, follow us on twitter @assertpub_ (arXiv), @assert_pub (bioRxiv), and @assertpub_dev (everything else).

To see beautiful figures extracted from papers, follow us on Instagram.

Tracking 225,721 papers.

Search
Sort results based on if they are interesting or reproducible.
Interesting
Reproducible
Categories
All
Astrophysics
Cosmology and Nongalactic Astrophysics
Earth and Planetary Astrophysics
Astrophysics of Galaxies
High Energy Astrophysical Phenomena
Instrumentation and Methods for Astrophysics
Solar and Stellar Astrophysics
Condensed Matter
Disordered Systems and Neural Networks
Mesoscale and Nanoscale Physics
Materials Science
Other Condensed Matter
Quantum Gases
Soft Condensed Matter
Statistical Mechanics
Strongly Correlated Electrons
Superconductivity
Computer Science
Artificial Intelligence
Hardware Architecture
Computational Complexity
Computational Engineering, Finance, and Science
Computational Geometry
Computation and Language
Cryptography and Security
Computer Vision and Pattern Recognition
Computers and Society
Databases
Distributed, Parallel, and Cluster Computing
Digital Libraries
Discrete Mathematics
Data Structures and Algorithms
Emerging Technologies
Formal Languages and Automata Theory
General Literature
Graphics
Computer Science and Game Theory
Human-Computer Interaction
Information Retrieval
Information Theory
Machine Learning
Logic in Computer Science
Multiagent Systems
Multimedia
Mathematical Software
Numerical Analysis
Neural and Evolutionary Computing
Networking and Internet Architecture
Other Computer Science
Operating Systems
Performance
Programming Languages
Robotics
Symbolic Computation
Sound
Software Engineering
Social and Information Networks
Systems and Control
Economics
Econometrics
General Economics
Theoretical Economics
Electrical Engineering and Systems Science
Audio and Speech Processing
Image and Video Processing
Signal Processing
General Relativity and Quantum Cosmology
General Relativity and Quantum Cosmology
High Energy Physics - Experiment
High Energy Physics - Experiment
High Energy Physics - Lattice
High Energy Physics - Lattice
High Energy Physics - Phenomenology
High Energy Physics - Phenomenology
High Energy Physics - Theory
High Energy Physics - Theory
Mathematics
Commutative Algebra
Algebraic Geometry
Analysis of PDEs
Algebraic Topology
Classical Analysis and ODEs
Combinatorics
Category Theory
Complex Variables
Differential Geometry
Dynamical Systems
Functional Analysis
General Mathematics
General Topology
Group Theory
Geometric Topology
History and Overview
Information Theory
K-Theory and Homology
Logic
Metric Geometry
Mathematical Physics
Numerical Analysis
Number Theory
Operator Algebras
Optimization and Control
Probability
Quantum Algebra
Rings and Algebras
Representation Theory
Symplectic Geometry
Spectral Theory
Statistics Theory
Mathematical Physics
Mathematical Physics
Nonlinear Sciences
Adaptation and Self-Organizing Systems
Chaotic Dynamics
Cellular Automata and Lattice Gases
Pattern Formation and Solitons
Exactly Solvable and Integrable Systems
Nuclear Experiment
Nuclear Experiment
Nuclear Theory
Nuclear Theory
Physics
Accelerator Physics
Atmospheric and Oceanic Physics
Applied Physics
Atomic and Molecular Clusters
Atomic Physics
Biological Physics
Chemical Physics
Classical Physics
Computational Physics
Data Analysis, Statistics and Probability
Physics Education
Fluid Dynamics
General Physics
Geophysics
History and Philosophy of Physics
Instrumentation and Detectors
Medical Physics
Optics
Plasma Physics
Popular Physics
Physics and Society
Space Physics
Quantitative Biology
Biomolecules
Cell Behavior
Genomics
Molecular Networks
Neurons and Cognition
Other Quantitative Biology
Populations and Evolution
Quantitative Methods
Subcellular Processes
Tissues and Organs
Quantitative Finance
Computational Finance
Economics
General Finance
Mathematical Finance
Portfolio Management
Pricing of Securities
Risk Management
Statistical Finance
Trading and Market Microstructure
Quantum Physics
Quantum Physics
Statistics
Applications
Computation
Methodology
Machine Learning
Other Statistics
Statistics Theory
Feedback
Online
Stats
Tracking 225,721 papers.