Adaptation and Evaluation of Recommendationsfor Short-term Shopping GoalsLukasLerche
Slides of the paper presentation in session 5b at the RecSys 2015 conference in Vienna on Friday, September 18th 2015.
Speaker:
Lukas Lerche, TU Dortmund, Germany
ACM link to the paper:
http://dl.acm.org/citation.cfm?id=2792838.2800176
Direct link to the paper:
http://ls13-www.cs.tu-dortmund.de/homepage/publications/jannach/Conference_RecSys_2015_st.pdf
Abstract:
An essential characteristic in many e-commerce settings is that website visitors can have very specific short-term shopping goals when they browse the site. Relying solely on long-term user models that are pre-trained on historical data can therefore be insufficient for a suitable next-basket recommendation. Simple "real-time" recommendation approaches based, e.g., on unpersonalized co-occurrence patterns, on the other hand do not fully exploit the available information about the user's long-term preference profile.
In this work, we aim to explore and quantify the effectiveness of using and combining long-term models and short-term adaptation strategies. We conducted an empirical evaluation based on a novel evaluation design and two real-world datasets. The results indicate that maintaining short-term content-based and recency-based profiles of the visitors can lead to significant accuracy increases. At the same time, the experiments show that the choice of the algorithm for learning the long-term preferences is particularly important at the beginning of new shopping sessions.
Adaptation and Evaluation of Recommendationsfor Short-term Shopping GoalsLukasLerche
Slides of the paper presentation in session 5b at the RecSys 2015 conference in Vienna on Friday, September 18th 2015.
Speaker:
Lukas Lerche, TU Dortmund, Germany
ACM link to the paper:
http://dl.acm.org/citation.cfm?id=2792838.2800176
Direct link to the paper:
http://ls13-www.cs.tu-dortmund.de/homepage/publications/jannach/Conference_RecSys_2015_st.pdf
Abstract:
An essential characteristic in many e-commerce settings is that website visitors can have very specific short-term shopping goals when they browse the site. Relying solely on long-term user models that are pre-trained on historical data can therefore be insufficient for a suitable next-basket recommendation. Simple "real-time" recommendation approaches based, e.g., on unpersonalized co-occurrence patterns, on the other hand do not fully exploit the available information about the user's long-term preference profile.
In this work, we aim to explore and quantify the effectiveness of using and combining long-term models and short-term adaptation strategies. We conducted an empirical evaluation based on a novel evaluation design and two real-world datasets. The results indicate that maintaining short-term content-based and recency-based profiles of the visitors can lead to significant accuracy increases. At the same time, the experiments show that the choice of the algorithm for learning the long-term preferences is particularly important at the beginning of new shopping sessions.
ICML2018読み会: Overview of NLP / Adversarial AttacksMotoki Sato
ICML 2018読み会の資料.
Overview of NLP/ Adversarial Attacks
- Obfuscated gradients give a false sense of security: circumventing defenses to adversarial examples
- Synthesizing Robust Adversarial Examples
- Black-box Adversarial Attacks with Limited Queries and Information
Productionizing Spark and the Spark Job ServerEvan Chan
You won't find this in many places - an overview of deploying, configuring, and running Apache Spark, including Mesos vs YARN vs Standalone clustering modes, useful config tuning parameters, and other tips from years of using Spark in production. Also, learn about the Spark Job Server and how it can help your organization deploy Spark as a RESTful service, track Spark jobs, and enable fast queries (including SQL!) of cached RDDs.
LinkedIn talk at Netflix ML Platform meetup Sep 2019Faisal Siddiqi
In this talk at the Netflix Machine Learning Platform Meetup on 12 Sep 2019, Kinjal Basu from LinkedIn discussed Online Parameter Selection for web-based Ranking vis Bayesian Optimization
Recsys 2015: Making Meaningful Restaurant Recommendations at OpenTableSudeep Das, Ph.D.
At OpenTable, recommendations play a key role in connecting diners with restaurants. The act of recommending a restaurant to a diner relies heavily on aligning everything we know about the restaurant with everything we can infer about the diner. Our methods go beyond using the diner-restaurant interaction history as the sole input — we use click and search data, the metadata of restaurants, as well as insights gleaned from reviews, together with any contextual information to make meaningful recommendations. In this talk, I will highlight the main aspects of our recommendation stack built with Scala using Apache Spark.
ICML2018読み会: Overview of NLP / Adversarial AttacksMotoki Sato
ICML 2018読み会の資料.
Overview of NLP/ Adversarial Attacks
- Obfuscated gradients give a false sense of security: circumventing defenses to adversarial examples
- Synthesizing Robust Adversarial Examples
- Black-box Adversarial Attacks with Limited Queries and Information
Productionizing Spark and the Spark Job ServerEvan Chan
You won't find this in many places - an overview of deploying, configuring, and running Apache Spark, including Mesos vs YARN vs Standalone clustering modes, useful config tuning parameters, and other tips from years of using Spark in production. Also, learn about the Spark Job Server and how it can help your organization deploy Spark as a RESTful service, track Spark jobs, and enable fast queries (including SQL!) of cached RDDs.
LinkedIn talk at Netflix ML Platform meetup Sep 2019Faisal Siddiqi
In this talk at the Netflix Machine Learning Platform Meetup on 12 Sep 2019, Kinjal Basu from LinkedIn discussed Online Parameter Selection for web-based Ranking vis Bayesian Optimization
Recsys 2015: Making Meaningful Restaurant Recommendations at OpenTableSudeep Das, Ph.D.
At OpenTable, recommendations play a key role in connecting diners with restaurants. The act of recommending a restaurant to a diner relies heavily on aligning everything we know about the restaurant with everything we can infer about the diner. Our methods go beyond using the diner-restaurant interaction history as the sole input — we use click and search data, the metadata of restaurants, as well as insights gleaned from reviews, together with any contextual information to make meaningful recommendations. In this talk, I will highlight the main aspects of our recommendation stack built with Scala using Apache Spark.
Rodina Márie Terézie,
Mária Terézia 1740-1780, Maria Theresa 1740-1780,
cisárovná Svätej ríše rímskej, Kráľovná Uhorska, Kráľovná Čiech, Arcivojvodkyňa Rakúska, dejiny, dejepis, history
Reformy Márie Terézie,
Mária Terézia 1740-1780, Maria Theresa 1740-1780,
cisárovná Svätej ríše rímskej, Kráľovná Uhorska, Kráľovná Čiech, Arcivojvodkyňa Rakúska, dejiny, dejepis, history
2. Ľudovo sa nazýva aj svätojánska
muška.
• Svetlušky si ľudové pomenovanie vyslúžili
aj preto, že ich je najviac vidieť okolo
sviatku sv. Jána (24. jún), no
zahliadnúť ich možno počas celého leta.
7. 3 časti tela: hlava, hruď, bruško
6 nôh
2 tykadlá
2 oči
Posledná časť bruška svieti
2 páry krídel = dospelí samčekovia
Samičky zvyčajne nemajú krídla a ani
Larvy.
8.
9. Ako svietia
• Na chvostíku = na spodnej strane šiesteho
a siedmeho článku majú „lampášik“.
• Ten obsahuje farbivo = pigment zvaný
luciferín
• Ich chvostík prijíma kyslík a pod vplyvom
kyslíka sa luciferín premieňa na svetlo.
• Ich svetlo je studené
10. Najúspornejšie svetlo na
svete
• Svetelný orgán svetlušiek je
mnohonásobne efektívnejší ako
najúspornejšie žiarivky vytvorené ľuďmi.
Až 98 percent uvoľnenej energie dokáže
premeniť na svetlo, kým ľudská žiarivka
iba 10 percent.
12. Komunikácia
• Samičky svietia viac, a to preto, aby
prilákali samčekov. Vyzývajú ich tak k
páreniu.
• Samčekovia lietajú jeden až dva metre
nad zemou, ak uvidia svietiace samičky,
zletia k nim na zem.
13. • Každý druh svetlušiek má svoj vlastný
jedinečný vzor blikania.
• Keď sa samec svetlušky chce dorozumieť
so samičkou svetlušky, letí pri zemi,
pričom každých šesť sekúnd zabliká
svetlom.
• Keď je pri zemi, samica ľahšie zistí, či je z
rovnakého druhu ako ona. (Väčšina
samičiek nevie lietať.) Odpovedá na jeho
záblesky rozsvietením svetiel. Potom ju
samec nájde.
14. Odstrašenie predátorov.Ich krv je
horká
• Väčšina zvierat sa vyhýba ich konzumácii.
• Hoci svetlušky ľahko rozoznajú podľa ich
žiary, jedia ich len zriedka. Je to preto, že
svetlušky uvoľňujú kvapky toxickej,
nechutnej krvi. Ich blikanie je varovným
svetlom pre dravcov, aby sa držali ďalej.
15. Potrava
Svetlušky sú mäsožravé:
Slimáky a húsenice. Jedia hlavne larvy, ktoré žijú 2 roky. Stravujú
sa zdravšie, ako dospelé jedince. Môže sa stať, že dospelá
svetluška ani neje, občas sa zastaví po nektár alebo peľ.
Potrebujú vodu.
16. Krátky život svetlušiek
• Dospelé svetlušky žijú len niekoľko dní – 2
týždne. Splodia nové potomstvo a potom
uhynú.
19. Larva 2 roky
• Nesvetielkujú iba dospelé jedince, ale aj
ostatné vývojové štádiá. Prevažnú časť
života prežije svetluška v podobe dravej
larvy.
• Larválne štádium svetielkujúceho hmyzu
totiž trvá až dva roky, počas ktorých
prežívajú ukryté v pôde alebo vo vode.
• (Preto sú citlivé na znečistenie vody a
pôdy = svetlušky vymierajú)
22. Tri farby
• Existuje viacero
druhov, tie na našom
území svietia
žltozeleno, tropické
na oranžovo a
japonský druh aj
modro.
23.
24.
25. Príroda zhasína
• svetlušky čelia reálnej hrozbe vyhynutia:
1. strata domova
2. používanie pesticídov (znečistenie pôdy a
vody, v ktorých žijú larvy 2 roky)
3. umelé svetlo ľudí narúša ich rituály
párenia
26.
27. Pomáha lekárom
• Vedci zvyčajne používajú luciferázu
svetlušiek na lekársky výskum. Môže sa
použiť na monitorovanie hladín peroxidu
vodíka, detekciu krvných zrazenín a
označenie buniek vírusu tuberkulózy.