[B! machine_learning] moozã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

mooz id:mooz

machine_learningã«é–¢ã™ã‚‹moozã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (277)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

Jupyter ã ã‘ã§æ©Ÿæ¢°å¦ç¿’ã‚’å®Ÿã‚µãƒ¼ãƒ“ã‚¹å±•é–‹ã§ãã‚‹åŸºç›¤ - Tech Blog - Recruit Lifestyle Engineer
ã“ã‚“ã«ã¡ã¯ã€CET ãƒãƒ¼ãƒ ã®ç”°æ‘ã§ã™ã€‚ãƒ‡ãƒ¼ã‚¿åŸºç›¤ã‚’æ§‹ç¯‰ãƒ»é‹ç”¨ã—ãŸã‚Šã€ãƒãƒ£ãƒƒãƒˆãƒœãƒƒãƒˆï¼ˆã‚¹ãƒžãƒ›ç”¨ã§ã™ï¼‰ã‚’é–‹ç™ºã—ãŸã‚Šã—ã¦ã„ã‚‹ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã§ã™ã€‚ çš†ã•ã‚“ã€å®Ÿã‚µãƒ¼ãƒ“ã‚¹ã§æ©Ÿæ¢°å¦ç¿’ã€æ´»ç”¨ã§ãã¦ã„ã¾ã™ã‹ï¼Ÿ æ£ç›´ã€é›£ã—ã„ã§ã™ã‚ˆãã€‚é«˜ç²¾åº¦ãªãƒ¢ãƒ‡ãƒ«ã‚’ä½œã‚‹ã“ã¨è‡ªä½“ã‚‚é›£ã—ã„ã§ã™ãŒã€å®Ÿéš›ã®ã‚µãƒ¼ãƒ“ã‚¹ã«ãã‚Œã‚’çµ„ã¿è¾¼ã‚€ã«ã¯ã€ãã“ã‹ã‚‰ã•ã‚‰ã«æ•°ã€…ã®é›£æ‰€ãŒå¾…ã¡æ§‹ãˆã¦ã„ã‚‹ã‹ã‚‰ã§ã™ã€‚ ã§ã‚‚ã€ãã®ã»ã¨ã‚“ã©ã¯ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ãƒªãƒ³ã‚°ã§è§£æ¶ˆã§ãã¾ã™ã€‚ ç§ãŸã¡ã®ãƒãƒ¼ãƒ ã§ã¯ã€æ•°å¹´ã«ã‚ãŸã‚‹çµŒé¨“ã‚’ã‚‚ã¨ã«é›£æ‰€ã¨ãã®å¯¾å‡¦æ³•ã‚’æ•´ç†ã—ã€ã™ã°ã‚„ãæˆæžœã‚’ã‚ã’ã‚‰ã‚Œã‚‹æ©Ÿæ¢°å¦ç¿’åŸºç›¤ã‚’é–‹ç™ºã—ã¯ã˜ã‚ã¾ã—ãŸã€‚ æœ¬è¨˜äº‹ã§ã¯ã€ã“ã®åŸºç›¤ã®è¨è¨ˆã¨ãã®èƒŒå¾Œã«ã‚ã‚‹ã‚¢ã‚¤ãƒ‡ã‚¢ã‚’ã”ç´¹ä»‹ã—ã¾ã™ï¼ˆæ©Ÿæ¢°å¦ç¿’å·¥å¦ç ”ç©¶ä¼šã®å‹‰å¼·ä¼šã§ã®ç™ºè¡¨è³‡æ–™ãŒãƒ™ãƒ¼ã‚¹ã§ã™ï¼‰ã€‚ ã‚¤ãƒ†ãƒ¬ãƒ¼ã‚·ãƒ§ãƒ³ã‚’ä½•åº¦ã‚‚å›žã› åŸºç›¤ãã®ã‚‚ã®ã®å‰ã«ã€ã¾ãšæ©Ÿæ¢°å¦ç¿’ã‚’æˆæžœã«ã¤ãªã’ã‚‹ãŸã‚ã®ãƒã‚¤ãƒ³ãƒˆã‚’èª¬æ˜Žã•ã›ã¦ãã ã•ã„ã€‚ ç§ãŸã¡ã¯ã€æ©Ÿæ¢°å¦ç¿’ã®æ´»ç”¨ã«ãŠã„ã¦å¿…è¦ãª
mooz 2018/10/04
è€ƒãˆæŠœã‹ã‚Œã¦ã„ã‚‹ã€‚MLOpsã€‚

machine_learning

mlops
ãƒªãƒ³ã‚¯
ã‚°ãƒ©ãƒ•ã‚«ãƒ¼ãƒãƒ«ã®ãƒ¡ãƒ¢
Weisfeiler-Lehmanã‚°ãƒ©ãƒ•ã‚«ãƒ¼ãƒãƒ«ã«ã¤ã„ã¦ã®ãƒ¡ãƒ¢ã§ã‚ã‚‹ï¼Žç‰¹ã«å†’é ã®ã‚°ãƒ©ãƒ•ã‚«ãƒ¼ãƒãƒ«ã‚’ä¿¯çž°ã—ãŸã‚¤ãƒ³ãƒˆãƒã‚ãŸã‚Šã‚’èªã‚“ã§é¢ç™½ã‹ã£ãŸã®ã§åŠ›å°½ãã‚‹ã¾ã§èªã‚“ã ï¼Žã‚ã¨æ›¸ã„ã¦ã„ã‚‹äººã¯ã‚«ãƒ¼ãƒãƒ«æ³•ã®ç ”ç©¶è€…ã§ã‚‚ãªã‘ã‚Œã°ï¼Œã‚°ãƒ©ãƒ•ã«é–¢ã—ãŸæ©Ÿæ¢°å¦ç¿’ã®ç ”ç©¶è€…ã§ã‚‚ã‚ã‚Šã¾ã›ã‚“ã®ã§ï¼Œæ°—è»½ã«èªã‚“ã§ãã ã•ã„ã(ã“ã‚Œã¯ã„ã‚ã‚†ã‚‹ï¼Œãƒžã‚¯ãƒ‰ãƒŠãƒ«ãƒ‰ã§å¥³åé«˜ç”ŸãŒXXãƒ¡ã‚½ãƒƒãƒ‰)ï¼Ž N. Shervashidze et al., Weisfeiler-Lehman Graph Kernels, JMLR, vol. 12, pp. 2539-2561, 2011. ãŠç´„æŸ: Preliminary ãƒ©ãƒ™ãƒ«ä»˜ãã‚°ãƒ©ãƒ•G=(V, E, l)ï¼Œãƒ©ãƒ™ãƒ«ã¯è‡ªç„¶æ•°ã¨ã™ã‚‹ï¼Ž N(v)ã§vã®è¿‘å‚ã‚’è¡¨ã™ï¼Ž Walk: edgeã®åˆ—ã§ï¼ŒPath: ç•°ãªã‚‹ç¯€ç‚¹ã‚’é€šã‚‹ã‚ˆã†ãªWalkï¼ˆãŸã¶ã‚“ï¼‰ éƒ¨åˆ†æœ¨: ã‚µã‚¤ã‚¯ãƒ«ã‚’å«ã¾ãªã„éƒ¨åˆ†ã‚°ãƒ©ãƒ• éƒ¨åˆ†æœ¨ãƒ‘ã‚¿ãƒ¼ãƒ³: åŒã˜ç¯€
mooz 2018/08/18
â€œWeisfeiler-Lehman Testâ€

machine_learning

graph

kernel
ãƒªãƒ³ã‚¯
Learning Disentangled Representations
Learning Disentangled Representations: from Perception to ControlNIPS 2017 Workshop An important facet of human experience is our ability to break down what we observe and interact with, along characteristic lines. Visual scenes consist of separate objects, which may have different poses and identities within their category. In natural language, the syntax and semantics of a sentence can often be
mooz 2018/07/23
è¡¨ç¾å¦ç¿’ã§å¾—ã‚‰ã‚ŒãŸFeatureã®è§£é‡ˆãªã©ã«é–¢ã™ã‚‹WS

machine_learning
ãƒªãƒ³ã‚¯
https://evoai.shinyapps.io/evoai/
mooz 2018/07/18
machine_learning
ãƒªãƒ³ã‚¯
Glow: Better reversible generative models
An interactive demo of our model to manipulate attributes of your face, and blend with other faces Manipulating attributes of images of researchers Prafulla Dhariwal and Durk Kingma. The model isnâ€™t given attribute labels at training time, yet it learns a latent space where certain directions correspond to changes in attributes like beard density, age, hair color, and so on. Generative modeling is
mooz 2018/07/10
flow-based generative model. VAEã¨é•ã„è¿‘ä¼¼ãŒå…¥ã‚‰ãªã„

machine_learning
ãƒªãƒ³ã‚¯
A survey on Optimal Transport for Domain Adaptation
mooz 2018/06/11
æœ€é©è¼¸é€ã€‚ãƒ‰ãƒ¡ã‚¤ãƒ³é©å¿œã€‚

machine_learning
ãƒªãƒ³ã‚¯
ã€è¨˜äº‹æ›´æ–°ã€‘ç§ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ã€Œæ©Ÿæ¢°å¦ç¿’ã«ãŠã‘ã‚‹è§£é‡ˆæ€§ï¼ˆInterpretability in Machine Learningï¼‰ã€ | äººå·¥çŸ¥èƒ½å¦ä¼š (The Japanese Society for Artificial Intelligence)
Home Â» ãƒªã‚½ãƒ¼ã‚¹ Â» ç§ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ Â» ã€è¨˜äº‹æ›´æ–°ã€‘ç§ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ã€Œæ©Ÿæ¢°å¦ç¿’ã«ãŠã‘ã‚‹è§£é‡ˆæ€§ï¼ˆInterpretability in Machine Learningï¼‰ã€ æ©Ÿæ¢°å¦ç¿’ã«ãŠã‘ã‚‹è§£é‡ˆæ€§ (Interpretability in Machine Learning)åŽŸ è¡ï¼ˆå¤§é˜ªå¤§å¦ç”£æ¥ç§‘å¦ç ”ç©¶æ‰€ï¼‰ ã¯ã˜ã‚ã«è¿‘å¹´ã®äººå·¥çŸ¥èƒ½ æŠ€è¡“ã€ç‰¹ã«æ©Ÿæ¢°å¦ç¿’ã®ç™ºå±•ã«ä¼´ã„ã€ã“ã‚Œã‚‰ã®æŠ€è¡“ã¸ã®ç¤¾ä¼šçš„ãªæœŸå¾…ãŒé«˜ã¾ã£ã¦ã„ã‚‹ã€‚ ã—ã‹ã—ã€ã“ã®ã‚ˆã†ãªæœŸå¾…ã®é«˜ã¾ã‚Šã¨åŒæ™‚ã«ã“ã‚Œã‚‰æŠ€è¡“ã¸ã®ä¸å®‰ã‚‚é«˜ã¾ã£ã¦ã„ã‚‹ã€‚ ç‰¹ã«ã€æ·±å±¤å¦ç¿’ãƒ¢ãƒ‡ãƒ«ã‚’åˆã‚ã¨ã™ã‚‹æ©Ÿæ¢°å¦ç¿’ãƒ¢ãƒ‡ãƒ«ãŒè¤‡é›‘ãªãƒ–ãƒ©ãƒƒã‚¯ãƒœãƒƒã‚¯ã‚¹ã§ã‚ã‚‹ãŒã‚†ãˆã«å®‰æ˜“ã«ä¿¡é ¼ã§ããªã„ã¨ã™ã‚‹æ‡¸å¿µã®å£°ãŒä¸ŠãŒã‚Šåˆã‚ã¦ã„ã‚‹ã€‚ ã“ã‚Œã«å¯¾ã—ã€ç·å‹™çœã¯AIã®åˆ©ç”¨ã®ä¸€å±¤ã®å¢—é€²ã¨ãã‚Œã«ä¼´ã†ãƒªã‚¹ã‚¯ã®æŠ‘åˆ¶ã®ãŸã‚ã«ã€ŒAIé–‹ç™ºã‚¬ã‚¤ãƒ‰ãƒ©ã‚¤ãƒ³æ¡ˆã€[1]ã‚’2017å¹´ã«ç–å®šã—ãŸã€‚ ã“ã®ã‚¬ã‚¤ãƒ‰ãƒ©ã‚¤ãƒ³æ¡ˆã§ã¯ã€ä¸Šè¨˜ã®ã‚ˆã†ãªæ‡¸å¿µ
mooz 2018/05/01
ç´ æ™´ã‚‰ã—ã„ã¾ã¨ã‚ã

machine_learning
ãƒªãƒ³ã‚¯
ã€Œãƒ•ã‚§ã‚¤ã‚¯ãƒ‡ãƒ¼ã‚¿ã€ãŒã€äººå·¥çŸ¥èƒ½ã‚’ã‚‚ã£ã¨è³¢ãã™ã‚‹
mooz 2018/04/30
machine_learning
ãƒªãƒ³ã‚¯
FeatureAgglomeration
FeatureAgglomeration# class sklearn.cluster.FeatureAgglomeration(n_clusters=2, *, metric='euclidean', memory=None, connectivity=None, compute_full_tree='auto', linkage='ward', pooling_func=<function mean>, distance_threshold=None, compute_distances=False)[source]# Agglomerate features. Recursively merges pair of clusters of features. Read more in the User Guide. Parameters: n_clustersint or None,
mooz 2018/04/10
ç‰¹å¾´é‡ã‚’å‡é›†ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°ã—ãŸã‚‰è§£é‡ˆæ€§çš„ã«ã‚‚æ€§èƒ½çš„ã«ã‚‚è‰¯ã•ãã†ã¨æ€ã£ãŸã‚‰sklearnã«ãƒ¢ã‚¸ãƒ¥ãƒ¼ãƒ«ãŒã‚ã£ãŸ

machine_learning
ãƒªãƒ³ã‚¯
http://dlvm.org/
mooz 2018/02/23
Deep Learningå‘ã‘ã‚³ãƒ³ãƒ‘ã‚¤ãƒ©ã‚¤ãƒ³ãƒ•ãƒ©ã‚¹ãƒˆãƒ©ã‚¯ãƒãƒ£ã€‚IR ãŒ first-citizen ãªãƒ†ãƒ³ã‚½ãƒ«ã‚„ç·šå½¢ä»£æ•°æ¼”ç®—ãªã©ã‚’æŒã¤ã€‚ä½•æ•…ã‹Swiftå®Ÿè£…ã€‚

machine_learning
ãƒªãƒ³ã‚¯
A Survey Of Methods For Explaining Black Box Models
In the last years many accurate decision support systems have been constructed as black boxes, that is as systems that hide their internal logic to the user. This lack of explanation constitutes both a practical and an ethical issue. The literature reports many approaches aimed at overcoming this crucial weakness sometimes at the cost of scarifying accuracy for interpretability. The applications i
mooz 2018/02/10
machine_learning
ãƒªãƒ³ã‚¯
Learning Efficient Tensor Representations with Ring Structure Networks
mooz 2018/01/30
tensor ring

machine_learning
ãƒªãƒ³ã‚¯
Learning explanatory rules from noisy data
Research Learning explanatory rules from noisy data Published 29 January 2018 Authors Richard Evans, Edward Grefenstette Suppose you are playing football. The ball arrives at your feet, and you decide to pass it to the unmarked striker. What seems like one simple action requires two different kinds of thought. First, you recognise that there is a football at your feet. This recognition requires in
mooz 2018/01/30
symbolic program synthesis

programming_language

machine_learning

deep_learning
ãƒªãƒ³ã‚¯
Strongly Hierarchical Factorization Machines and ANOVA Kernel Regression
High-order parametric models that include terms for feature interactions are applied to various data mining tasks, where ground truth depends on interactions of features. However, with sparse data, the high- dimensional parameters for feature interactions often face three issues: expensive computation, difficulty in parameter estimation and lack of structure. Previous work has proposed approaches
mooz 2018/01/16
machine_learning
ãƒªãƒ³ã‚¯
The Unreasonable Effectiveness of Structure - NIPS 2017 Keynote - Lise Getoor
mooz 2018/01/05
probabilistic soft logic. ãƒ‡ãƒ¼ã‚¿ãŒã‚¹ãƒ‘ãƒ¼ã‚¹ã ã£ãŸã‚ŠãƒŽã‚¤ã‚¸ãƒ¼ã ã£ãŸã‚Šã™ã‚‹å ´åˆã¯æœ€è¿‘æµè¡Œã‚Šã®ã‚°ãƒ©ãƒ•åŸ‹ã‚è¾¼ã¿ã«ã‚ˆã‚‹é–¢ä¿‚æ€§äºˆæ¸¬ã‚ˆã‚Šé«˜ã„æ€§èƒ½ã¨ãªã‚ŠãŒã¡

machine_learning

logic
ãƒªãƒ³ã‚¯
Exponential Stochastic Cellular Automata for Massively Parallel Inference
mooz 2018/01/04
Alex Smola ã¨ Guy Steele ãŒå…±è‘—ã§ AISTATS ã«è«–æ–‡ã‚’å‡ºã—ã¦ã„ã‚‹ã®ã€èƒ¸ç†±ãªæ°—æŒã¡ã«ãªã‚‹ã€‚ã‚»ãƒ«ã‚ªãƒ¼ãƒˆãƒžãƒˆãƒ³ã«ã‚ˆã‚‹ç¢ºçŽ‡ãƒ¢ãƒ‡ãƒ«ã®æŽ¨è«–æ–¹å¼ã€‚

machine_learning

data_structure
ãƒªãƒ³ã‚¯
Reddit - Dive into anything
mooz 2017/12/29
ã‚¯ãƒ©ã‚¦ãƒ‰ãƒ™ãƒ¼ã‚¹ã®MLç³»ã®ãƒ—ãƒ©ãƒƒãƒˆãƒ•ã‚©ãƒ¼ãƒ

machine_learning
ãƒªãƒ³ã‚¯
InterSpeech2016
Being Shallow and Random is (almost) as Good as Being Deep and Thoughtful Fei Sha Joint work with collaborators from IBM and Columbia Kernel methods Motivation Deep neural networks Learning representation Motivating toy example: face recognition Dude Lady Step 1: collect labeled images as training data Dude Lady Step 2: represent each image as a point x1 x2 Step 3: fit a model (decision boundary)
mooz 2017/12/03
Random Fourier Feature ã§å¾—ã‚‰ã‚ŒãŸé«˜æ¬¡å…ƒãƒžãƒƒãƒ—ã‚’å˜ç´”åŒ–ã™ã‚‹è©±

kernel

machine_learning
ãƒªãƒ³ã‚¯
https://scikit-learn.org/stable/modules/sgd.html
mooz 2017/11/26
Empirically, we found that SGD converges after observing approx. 10^6 training samples. Thus, a reasonable first guess for the number of iterations is n_iter = np.ceil(10**6 / n), where n is the size of the training set.

machine_learning

mathematical_optimization
ãƒªãƒ³ã‚¯
CS231n Convolutional Neural Networks for Visual Recognition
Table of Contents: Gradient checks Sanity checks Babysitting the learning process Loss function Train/val accuracy Weights:Updates ratio Activation/Gradient distributions per layer Visualization Parameter updates First-order (SGD), momentum, Nesterov momentum Annealing the learning rate Second-order methods Per-parameter adaptive learning rates (Adagrad, RMSProp) Hyperparameter Optimization Evalua
mooz 2017/11/25
SGDã§ã®å¦ç¿’çŽ‡ã®èª¿ç¯€ã¨ã‹è‰²ã€…

machine_learning
ãƒªãƒ³ã‚¯
1 2 3 4 5 6 7 8 9 10 æ¬¡ã®ãƒšãƒ¼ã‚¸