Nov 2019
18 Mon
19 Tue
20 Wed
21 Thu
22 Fri
23 Sat 08:30 AM – 05:30 PM IST
24 Sun
Accepting submissions till 01 Nov 2019, 04:20 PM
Not accepting submissions
##Format:
Anthill Inside 2019 is a single track conference. Birds of Feather (BOF) sessions, round table discussions and office hours with speakers will be held in parallel with talks in the main auditorium.
We are accepting proposals for:
##Topics for submitting talks, workshops, tutorials and BOF sessions:
Practical application of concepts, including:
1.1 Bayesian Networks
1.2 Reinforcement learning
1.3 Knowledge Graphs
1.4 Interpretability
Data storage case studies: we are specifically interested in hearing about:
2.1 What do you do with data that has lost its currency?
2.2. How do you deal with privacy issues for vast amounts of irrelevant data?
Tools for AI, ML and Deep Learning. Here, we want to hear about:
3.1 Whether you use third-party tools? If yes, why?
3.2 Do you adopt and retrofit existing tools? Again, why? Give us a detailed case study.
3.3 Do you develop tools for AI/ML/DL in-house? Why are in-house tools necessary for your case?
Storing data on the cloud and cloud strategy.
4.1 Do you have a multi-cloud strategy? Share this with the community.
4.2 How do you deal with lock-in situations with single providers?
GPU versus CPU -- when you do use either and why? How do the strengths and limitations of each play out for your use case?
Case studies of practical applications of Computer Vision, NLP and Deep Learning to solve business problems.
##Anthill Inside’s speaking policies:
We only accept one speaker per talk. This is non-negotiable. Workshops or tutorials may have more than one instructor.
##Selection criteria:
The first filter for a proposal is whether the technology or solution you are referring to is open source or not. The following criteria apply for closed source talks:
The criteria for selecting proposals, in the order of importance, are:
##How to submit a proposal (and increase your chances of getting selected):
The following guidelines will help you in submitting a proposal:
##Passes and honorarium for speakers:
We pay an honararium of Rs. 3,000 to each speaker. Confirmed speakers also get a pass to the conference. We do not provide free passes for speakers’ colleagues and spouses. Please don’t ask us for this.
##Travel grants for outstation speakers:
Travel grants are available for international and domestic speakers. We evaluate each case on its merits, giving preference to women, people of non-binary gender, and Africans. If you require a grant, request it when you submit your proposal in the field where you add your location. Anthill Inside is funded through ticket purchases and sponsorships; travel grant budgets vary.
##Contact details:
For information about speaking or teaching workshops at Anthill Inside, write to anthillinside.editorial@hasgeek.com or call +91 7676332020
Accepting submissions till 01 Nov 2019, 04:20 PM
Virtual Assistant for High Volume RecruitmentLogistics companies, both old and new, have invested heavily in building an efficient frontline workforce to provide swift and convinient services to their users. Timely delivery is often a critical deciding factor for the ever-impatient customers to choose service A over service B. Hence, operations/logistic team is the key enabler here. more
Technical level: Intermediate
|
Portfolio Optimization using Deep Reinforcement LearningWhat is Portfolio Management? What is Deep Learning? How does one apply deep learning to the complex problem of portfolio management? What is the intuitive interpretation of this application? What lies under the black box? more
Section: Full talk
Technical level: Advanced
Session type: Demo
|
Production Object Detection - A Journey of Training, Building and Deploying CV modelsComputer Vision as a field has changed manifold in the past few years. Researchers publish their papers and at times their code for the latest algorithms, but the challenge for the industry remains in applying that research to their processes. Customising a company’s proprietary data for the research models, implementing their code, and training models is the first big hurdle. Then comes the part… more
Technical level: Beginner
|
Essential Python Recipes for Deep LearningThis coding-focused talk offers some practical tips and advice that can help beginners be more efficient, organized and productive while working on deep learning projects and participating in data science competitions. The topics covered include working with the file system, effecting processing CSV files, doing exploratory data analysis, creating a baseline model for evaluation, and easily worki… more
Section: Crisp talk
Technical level: Beginner
|
Exploring the un-conventional: End-to-End learning architectures for automatic speech recognitionSpeech recognition is a challenging area, where accuracies have risen dramatically with the use of deep learning over the last decade, but there are still many areas of improvement. We start with the basics of speech recognition and the design of a conventional speech recognition system, comprising of acoustic modeling, language modeling, lexicon (pronunciation model) and decoder. To improve the … more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
AgentBuddy: Leveraging Bandit Algorithms for a human-in-loop system for Customer Care Agents (Paper accepted for the demo track at SIGIR-2019)We have developed a human-in-the loop system, AgentBuddy that is helping Intuit improve the quality of search it offers to internal Customer Care Agents (CCAs). AgentBuddy aims to reduce the cognitive effort on part of the CCAs while at the same time boosting the quality of our legacy federated search system. It addresses two key pain points 1)Given several candidate query answering mechanisms, h… more
Section: Full talk
Technical level: Intermediate
Session type: Demo
|
Weaponising Artificial Intelligence In Cyber Security - The Next Age of Cyber Security EndgameMy talk discusses my research on how to integrate AI technologies into cyber security industry intelligently. more
Section: Crisp talk
Technical level: Intermediate
|
Non-Intent User Similarity for recommendation systemsIn the world of Ad Business’s recommendation systems it is easier comparatively to recommend to user who have shown some intent. But what about the users who have not shown any intent? How do you target them? In this talk I will like to talk about a novel approach to use user similarity from supply data to work out significant recommendation for these users more
Section: Full talk
Technical level: Intermediate
|
Snorkeling in the deep: Bootstrapping an NLU modelConsider building a natural language understanding model for powering task based conversational agents. One of the problems to be solved is slot extraction. For example, if a user utters “show me flights from bengaluru to delhi on 25th july”, the model needs to extract the slots {from: bengaluru, to: delhi, date: 25-07-2019}. Recent advances in deep learning can solve this problem with adequate t… more
Section: Crisp talk
Technical level: Beginner
Session type: Lecture
|
Attention based sequence to sequence models for natural language processing##Workshop details including schedule, venue, date and tickets are published here: https://hasgeek.com/anthillinside/sequence-to-sequence-models-workshop/ more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Productionizing deep learning workflow with Hangar, <frameworkOfYourChoice> & RedisAIManaging DL workflow is always a nightmare. Problems include handling the scale, efficient resource utilization, version controlling the data etc. With the heavily organized Hangar, we can keep the data on check now, not as a blob but as tensors in the data store and version at. The super flexible PyTorch gives us the advantage of prototyping faster and iterate smoother. The model prototype can n… more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Annotate This! A simple tool that seeks to simplify the creation of annotated text datasets - primarily for Hindi and other vernacular LanguagesThis talk will comprise of my giving a quick overview of an open source tool that allows for faster and relatively error free annotations of text data, by leveraging language models and gamification principles. I would subsequently go over the tech that powers it and demonstrate how it works. more
Section: Crisp talk
Technical level: Beginner
Session type: Demo
|
Tensorboard: Almost a one stop shop for Machine Learning DevelopmentThis talk will focus on the areas of Machine Learning development (specifically Computer Vision problems) and will discuss how using certain tools will make your life easier. Also, discuss some areas where developing custom tools is beneficial instead of using open-source tools and the trade-off for making this choice. We will discuss why we should be using Tensorboard and why it is the only tool… more
Section: Full talk
Technical level: Beginner
Session type: Lecture
|
Hands on Deep Learning for Computer Vision – Techniques for Image Segmentation.(6 hours workshop).Abstract – Computer Vision has lots of applications including medical imaging, autonomous vehicles, industrial inspection and augmented reality. Use of Deep Learning for computer Vision can be categorized into multiple categories for both images and videos – Classification, detection, segmentation & generation. Having worked in Deep Learning with a focus on Computer Vision have come across variou… more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Introduction to Bayesian Networks(3 hour workshop)Most machine learning models assume independent and identically distributed (i.i.d) data. Graphical models can capture almost arbitrarily rich dependency structures between variables. They encode conditional independence structure with graphs. Bayesian network, a type of graphical model describes a probability distribution among all variables by putting edges between the variable nodes, wherein e… more
Section: Workshops
Technical level: Beginner
Session type: Workshop
|
Reinforcement LearningWe see quite a lot of excitement around Reinforcement Learning (RL) in industry and academia. We regularly see new industrial applications both inside and outside of Amazon. We also see RL papers dominating many ML conferences such as NeurIPS and ICML. RL is promising for multiple reasons that include a better formulation of the problem which allows for optimization of multi-step decision making … more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Learning to Rank recommendation - Ranknet to LambdaMART to Groupwise scoring functions - experiments, introduction to Tensorflow RankingSearch and product recommendations are typically served using CF, MF, FM techniques, content/context/sequence based methods or using learning to rank framework which is more generic than rest. Evolving from traditional classification and regression modeling methods, the loss functions, gradients, computation tricks have evolved to suit ranking problems (point wise to pair wise to list wise soluti… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Out of Distribution Detection in Deep Learning ClassifiersA common problem when using deep neural network models for classification problems is handling out of distribution data. In such scenarios, the classifiers tend to assign the new data point to one of the known classes with high probability, which can lead to unintended and potentially harmful consequences. At MakeMyTrip (MMT), we use deep learning based NLP classifier for understanding intent of … more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Interpret-ability as a bridge from Insights to Intuition in Machine and Deep LearningRequisite characteristics of the Machine Learning models that make them fully deployable in a business setting are multivarious and sometimes compelling. Mere predictive power and validation accuracy are sometimes not sufficient. These models need to be interpretable, bias-free, transparent, explainable, consistent and should have a global and local basis for their output predictions. In this tal… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Explainable AI: Behind the ScenesWith the Ai based systems proliferating in to all applications of the daily life, getting a better insight of their working mechanism is a much sought-after affair. Often the rationale behind the decision thrown out by an AI system is not well understood. Which part or feature of the input has influenced the decision to what extent is not known. This presentation provides insights to demystify th… more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
Feature selection and engineering using genetic algorithms and genetic programmingWhile feature selection is almost a solved problem in data science, feature engineering is still quite a mystery. In this talk I will outline a method that I use to solve feature engineering, with a goal to provide a generalized framework to tackle both feature engineering and selection simultaneoously. more
Section: Full talk
Technical level: Advanced
Session type: Lecture
|
Tools for AI & ML for machine learning at Scale.As applications become more business critical and application teams are receiving monitoring data for these mission critical business applications as a continuous stream it becomes difficult to manually monitor them and create dashboards/reporting around these applications. more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Using AI for improving performance and design of ad creativesThe ad creative is the image of the ad that the user sees on their device. It includes information such as the image dimensions, its placement on the screen or within the app, the native resolution of the creative and the mobile screen it is displayed on and similar parameters. Why exactly a particular ad is visually appealing and another not is hard to determine. Using data from the performance … more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Using Locations for Online-Behaviour Prediction with Sparse DataUser behaviour varies with their current location, as a consequence their engagement with online media (say Ads) varies with where they are; Knowing the type of location can help us target the user better and recommend better. more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Recommendation @ ScaleRecommendation is one of the most traditional and wide spread use case of Machine Learning. In this talk we want to showcase, how an advanced recommendation engine can be served at scale in Glance. Glance is an AI-powered, content driven, personalised Screen Zero (lockscreen) platform for mobile, which is used by over 26M DAU users in India. The talk will take you through each component of a reco… more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Opening the Black Box: How to Interpret Machine Learning models; techniques, tools, and takeawaysInterpretability of a model is the degree to which a human can consistently predict the model’s result. The higher the interpretability of a machine learning model, the easier it is to comprehend why certain decisions or predictions have been made. While interpretability is not important in low-risk domains and black box models abound, in domains such as medicine or finance, or high-risk domains … more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Bridging the gap between research to the deployment of Machine Learning modelsIn this talk, we propose a way to minimize the effort required to move data munging, visualization and training for ML models into production. Python notebooks are a go-to tools to carry out such research. Moving from this painstaking research to a full blown deployment requires too much effort and care from a data scientist’s perspective. Attempting to make the modeling process amiable to produc… more
Section: Full talk
Technical level: Intermediate
Session type: Workshop
|
Modeling the effects of blurriness in mobile adsThe creative is the image of the ad that the user sees and engages with upon viewing. This talk studies the effect of an ad creative’s specifications and quality of render in performance campaigns and suggests a playbook for digital marketers based on the findings and insights. An offline study on an ad creative’s specifications such as resolution, aspect ratio, handset density, device orientatio… more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Large scale Machine Learning and data storage for CDP: transforming Digital MarketingWe will talk about why do we need a single end-to-end customer data platform to enable truly personalised digital marketing. We also explain what pain-points, such as cold-start problem, do we solve for marketers if we collate and utilise data from first, second and third party sources rather than relying on just first party data. Then we will focus on the motive to use Machine Learning to create… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
End to End Computer Vision paradigm with respect to advanced deep learning techniques.Deep learning based approaches to solve image classification have become a core technology in AI, largely due to developments in computing powers and digital data. However image classification gained popularity beyond academic circle with the advent of visual object recognition challenge. more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
GAN-inspired Innovations in Computer Vision"The most interesting idea in the last 10 years in ML.” - Yann LeCun, Facebook AI research director. more
Section: Tutorials
Technical level: Intermediate
Session type: Lecture
|
Practical Recommendation Systems: Scalability, Accuracy, LatencyThis tutorial shall cover traditional and modern recommendation systems from a perspective of practical application, in an easy question-answer format. The content of this tutorial is derived from multiple state-of-the-art research papers as well as classical text books on recommendation systems. more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
Taking AI Products to MarketWant to have specific focussed discussion on what it takes to take pure play AI products to market. Have first hand experience taking multiple AI products to Indian B2B Market, I wanted to get product managers and KBLs in the domain to come in and share their journey and experiences. Hence, suggesting a bird of feather session. more
Section: Birds Of Feather (BOF) session
Technical level: Intermediate
Session type: Lecture
|
Dataset Denoising : Improving Accuracy of NLP ClassifierReliable evaluation for the performance of classifiers depends on the quality of the data sets on which they are tested. During the collecting and recording of a data set, however, some noise may be introduced into the data, especially in various real-world environments, which can degrade the quality of the data set. In this talk we will discuss how we at MakeMyTrip are continuously improving per… more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Use cases of Financial Data Science Techniques in retailFinancial domains like Insurance and Banking have uncertainty itself as an inherent product feature, and hence makes extensive use of Statistical models to develop, valuate and price their products. This presentation will showcase some of the techniques like Survival models and cashflow prediction models, popularly used in financial products, how can they be used in Retail data science, by showca… more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
Email Data AnalyticsSentiment Analysis from text is a well known problem in machine learning where a given text document can be either positive, negative or neutral. In the last few years, Sentiment Analysis has become a hot-trend topic of scientific and market research in the field of Natural Language Processing (NLP) and Machine Learning. more
Session type: Lecture
|
Building a Context-Aware Knowledge graph using Graph analysis & Language modelsIntroduction At EtherLabs, we are building a video platform that provides insights into the numerous live audio and video meetings that organizations conduct everyday. In such a scenario, in order to acquire critical metrics such as important moments, topics discussed and possible intents from textual data, basic NLP tasks like keyphrase extraction becomes significantly important. Important keyph… more
|
Network health predictions and optimization recommendation using Deep learning Neural network models and Reinforcement learningTime series prediction of network parameters and detecting network health with network performance optimization, has been an interesting problem to solve for researchers in the field of Machine Learning and Data Mining community. These use cases are present across different industries like retail, telecom, transport with good presence in Telecom industry. However, there remains a challenge in get… more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Introduction to Probabilistic Programming - PyMC3 and EdwardProbabilistic programming differ from deterministic ones by allowing language primitives to be stochastic. In other words, instead of being restricted to deterministic assignments such as: more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
Advanced NLP and Deep Learning for document classification - A case study in civil aviation safety prognosisIn this presentation, I apply a set of data-mining and sequential deep learning techniques to accident reports published by the National Transportation Safety Board (NTSB), in order to support real-time prognosis of adverse events. The focus here is on learning with text data that describes sequences of events. NTSB creates post-hoc investigation reports which contain raw text narratives of their… more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
Hacking Self-attention architectures to address Unsupervised text tasksSelf-attention architectures like BERT, OpenAI GPT, MT-DNN are current state-of-the art feature extractors for several supervised downstream tasks for text. However, their ability on unsupervised tasks like document/sentence similarity are inconclusive. In this talk, I intend to cover brief overview of self attention architectures for Language Modelling, fine-tuning/feature selection approaches f… more
|
Industrialized Capsule Networks for Text AnalyticsMulti-label text classification is an interesting problem where multiple tags or categories may have to be associated with the given text/documents. Multi-label text classification occurs in numerous real-world scenarios, for instance, in news categorization and in bioinformatics (gene classification problem, see [Zafer Barutcuoglu et. al 2006]). Kaggle data set is representative of the problem: … more
|
NLP bootcampRecent advances in machine learning have rekindled the quest to build machines that can interact with outside environment like we human do - using visual clues, voice and text. An important piece of this trilogy are systems that can process and understand text in order to automate various workflows such as chat bots, named entity recognition, machine translation, information extraction, summariza… more
Session type: Workshop
|
A Guide on Dynamic Parameter Estimation for Causal ForecastingExecutives now-a-days rely on forecasting approcahes in virtually any decision making. The use cases are ubiquitous in business and technology domains ranging from Demand/Sales forecasting in Supply Chain Management, Hiring/Attrition rate forecasting in HR operations, Predicting the cell traffic or netwrok health parameters for a Telecom business, just to name a few. The challenges faced in forec… more
Session type: Lecture
|
Accountable Behavioural Change Detection (VEDAR) using Machine LearningWith exponential increase in the availability of telemetry / streaming / real-time data, understanding contextual behavior changes is a vital functionality in order to deliver unrivalled customer experience and build high performance and high availability systems. Real-time behavior change detection finds a use case in number of domains such as social networks, network traffic monitoring, ad exch… more
Session type: Short talk of 20 mins
|
Bandit algorithms to Reduce Cognitive Load on Customer Care Agents (Paper accepted for the demo track at SIGIR-2019)We will describe a human-in-the loop system, AgentBuddy that is helping Intuit improve the quality of search it offers to internal Customer Care Agents (CCAs). AgentBuddy aims to reduce the cognitive effort on part of the CCAs while at the same time boosting the quality of our legacy federated search system. It addresses two key pain points 1)Given several candidate query answering mechanisms, ho… more
Session type: Full talk of 40 mins
|
Generation of Newsletter using Natural Language GenerationWe will generate event narratives using NLG. Assume I am part of event (like Anthill Inside). I conducted a session on particular topic (NLG). I make entry into internal tool. Below information are captured in database. more
Section: Crisp talk
Technical level: Beginner
Session type: Lecture
|
Learning to Rank framework for product recommendation - Ranknet to LambdaMART to Groupwise scoring functions - experimentsSearch and product recommendations are typically served using CF, MF, FM techniques, content/context/sequence based methods or using learning to rank framework which is more generic than rest. Evolving from traditional classification and regression modeling methods, the loss functions, gradients, computation tricks have evolved to suit ranking problems (point wise to pair wise to list wise soluti… more
Session type: Full talk of 40 mins
|
Design a real-time anomaly detection application using Spark and Machine LearningOur team works on a streaming application that produces outputs in the form of real-time comparisons between us and our competitors. The comparisons are consumed by our market managers and hotel partners and are used by them for making their day-to-day decisions and prioritize their business actions for the day. Owing to the importance and scale of the data, it is important to ensure that our Str… more
Section: Crisp talk
Technical level: Beginner
Session type: Discussion
|
Building a time series model using CNNs and GANsTime series anomaly detection and classification problems have existed and there are various existing solutions to tackle such kind of problems. However, all/most of the solutions are for ideal cases (having enough labelled data or developing a model with average precision/recall) and does not take into account the practical constraints of implementing and deploying a highly generalizable solutio… more
Session type: Short talk of 20 mins
|
Time Series Anomaly detection on structured data from IOT Network using CNNTime series anomaly detection and classification problems have existed and there are various existing solutions to tackle such kind of problems. However, all/most of the solutions are for ideal cases (having enough labelled data or developing a model with average precision/recall) and does not take into account the practical constraints of implementing and deploying a highly generalizable solutio… more
Session type: Short talk of 20 mins
|
Machine Learning Model and Dataset VersioningIn this talk we will discuss about the current practices of organizing ML projects using traditional open-source tool set like Git and Git-LFS as well as this tool set limitation. Thereby motivation for developing new ML specific version control systems will be explained. more
Section: Crisp talk
Technical level: Intermediate
Session type: Lecture
|
iCASSTLE: Imbalanced Classification Algorithm for Semi Supervised Text LearningInformation in the form of text can be found in abundance in the web today, which can be mined to solve multifarious problems. Customer reviews, for instance, flow in across multiple sources in thousands per day which can be leveraged to obtain several insights. Our goal is to extract cases of a rare event e.g., recall of products, allegations of ethics or, legal concerns or, threats to product-s… more
Section: Full talk
Technical level: Advanced
Session type: Lecture
Session type: Short talk of 20 mins
|
Using NLP to generate QuizzesThe Problem Quizzes/Questions are required by students for self-evaluation and by teachers to test students. more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Yes! Attention is all you need for NLPNatural language processing a very tough problem to crack. We humans have our own style of speaking and though many a times we might mean the same thing we say it differently. This makes it very difficult for the machine to understand and process language at human level. more
Session type: Lecture
Session type: Full talk of 40 mins
|
Birds of a Feather on InterpretabilityComplex machine learning models work very well at prediction and classification tasks but become really hard to interpret. On the other hand simpler models are easier to interpret but less accurate and hence oftentimes we are made to take a call between interpretability and accuracy. more
Section: Birds Of Feather (BOF) session
Technical level: Intermediate
Session type: Discussion
|
Myths and Realities of Data Labeling for Deep LearningIn this BoF, we will explore data labeling tasks for NLP and CV problems. Specifically, we will discusses nuiances around defining, crowd sourcing and executing data labeling tasks, along with quality assurance processes. We shall also discuss machine aided data taggint to save cost, time and efforts on different data labeling tasks. Finally, we shall also touch upon feedback loopswhen some of th… more
Section: Birds Of Feather (BOF) session
Technical level: Intermediate
Session type: Discussion
|
Demystifying deep reinforcement learning##Details (date, time, venue) and tickets for this workshop are available here: https://hasgeek.com/anthillinside/deep-reinforcement-learning-workshop/ more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
PySpark for GeoSpatial DataGeoSpatial data is the key data source when it comes to external data sources but this data is often too large to process. This is where PySpark comes in to reduce the computation time and makes the whole code more than 5 times faster. This workshop aims to solve the problem of calculating land covered with greenry of a region (Delhi) using Satellite images and Python. more
Section: Tutorials
Technical level: Beginner
Session type: Tutorial
|
Building Products with ML: A Workshop for Product & Engg ManagersMachine learning (ML) has seen substantial adoption, and a large number of data science teams are being created. Taking on ML projects requires product managers and engineers to learn an ML approach to problem solving, to be able to effectively work with data scientists and data engineers. There exists a huge gap in understanding more
Section: Workshops
Technical level: Intermediate
Session type: Workshop
|
Feast: Feature Store for Machine LearningFeatures are key to driving impact with AI at all scales, allowing organizations to dramatically accelerate innovation and time to market. Willem Pienaar explain how GOJEK, Indonesia’s first billion-dollar startup, unlocked insights in AI by building a feature store called Feast, and some of the lessons they learned along the way. more
Section: Full talk
Technical level: Beginner
Session type: Lecture
|
Document digitization - Rethinking it with Deep LearningWhen you think about Document digitisation from a business optimization process perspective, just performing OCR does not truly solve the problem. We at omni:us are building AI systems to support the insurance industry by handling claims. In order to achieve this we are performing various human-esque activities on so many different types of documents like page / document classification, informati… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Rigorous Evaluation of NLP Models for Real World DeploymentRapid progress in NLP Research has seen a swift translation to real world commercial deployment. While a number of success stories of NLP applications have emerged, failures of translating scientific progress in NLP to real-world software have also been considerable (some of these issues are covered in my IJCAI paper https://www.ijcai.org/proceedings/2018/717). Specifically, the challenges and ga… more
Section: Full talk
Technical level: Intermediate
Session type: Discussion
|
Tutorial on Testing of Machine Learning Applications##URL for workshop date, time, venue, schedule and tickets: https://hasgeek.com/anthillinside/testing-machine-learning-applications-workshop/ more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
Hangar; git for your dataSoftware development is entering an era where the behavior of programs critically depends on the data they were trained on. In this setting, data is the new source code, and this opens the door to challenges like versioning and collaboration on numerical data. Enter Hangar, an open-source tool by [tensor]werk that brings Git-style version control to n-dimensional arrays. It supports versioning, b… more
Section: Full talk
Technical level: Intermediate
Session type: Demo
|
Unsupervised Catalog Generation with Clustering, Reinforcement and MoreThis presentation will look at how you can generate product catalogs from ecommerce websites using just the homepage URL of the website. Techniques explored include URL clustering, regex generation, reinforcement learning and supervised classification. more
|
Probabilistic Modeling – a tutorial on Bayesian Networks##Details of the workshop (date, time, venue) and tickets available here: https://hasgeek.com/anthillinside/bayesian-networks-tutorial/ more
Section: Tutorials
Technical level: Intermediate
Session type: Tutorial
|
Building a Recommendation Engine for diverse content and user behaviorsGood recommendation systems, as we all know, are a great way to acquire users, creating a delightful user experience for user engagement while driving incremental revenue. There is a lot of innovation and research in making recommendations systems understand the user preferences and hence personalise better. more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
How we built our own scalable, real-time framework to create and serve conversational systems.Voice based systems are becoming pervasive in the world, and NLP systems that power these systems have the daunting task of understanding, with high precision and high coverage(recall), what a user says, and also, doing it in real time. There are many frameworks that help create these paths, and over time the building of conversational agents have been distilled into few primitives such as intent… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Artificial Intelligence for automated investmentNeotic.ai introduced the use of AI for assisting investors in decision making. The technology behind it is based on machine learning algorithms for price patterns recognition, corporate fundamentals, and financial news analysis. The company offers customized AI to hedge funds as well as plug and play, fully automated trading strategies. more
Section: Full talk
Technical level: Advanced
Session type: Lecture
|
Sizing biological cells and saving lives using AIAI techniques are finding applications in a wide range of applications.Crowd counting deep learning models have been used to count people, animals, and microscopic cells. This talk will introduce some novel crowd counting techniques and their applications. A pharma case study will be presented to show how it was used for drug discovery to bring about 98% savings in drug characterization efforts. more
Section: Birds Of Feather (BOF) session
Technical level: Intermediate
Session type: Demo
|
Model Interpretability, Explainable AI and the Right to InformationIssues of ‘explainability in AI’ have emerged as an important theme in the development of machine learning and statistical modelling. Most studies look at explainability through the lens of model interpretability, in order to understand underlying machine learning models better and improve them for better optimisation. However, there is limited relevance of this understanding of interpretability … more
Section: Crisp talk
Technical level: Beginner
Session type: Discussion
|
Blueprint for Building AI ProductsIT Product Development is changing at a rapid pace. From being an enabler of existing products, AI and ML has taken a Centerstage in Product development. more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
What can software learn from robots and mathIn almost every aspect of life, there is a concept of state (for example, where are you right now?) and a concept of uncertainty (for example, how soon can you reach home?). And whenever there is uncertainty, there is a tool that can help you get a better estimate of that state. It is called a Kalman Filter, named after an American Mathematician, Rudolf Kalman. more
Section: Full talk
Technical level: Advanced
Session type: Lecture
|
The shape of UTensors are the fundamental data structure for building modern machine learning programs and complex neural architectures. Unfortunately, the foundations of popular tensor libraries (numpy, tensorflow, pytorch) are hardly robust, e.g., tensor broadcasting rules are adhoc, and may cause surprising bugs. Further, the tensor library APIs expose low-level memory models to the developer, forcing them … more
Section: Crisp talk
Technical level: Advanced
Session type: Lecture
|
ML application lifecycle: recommendations for each stageBuilding good ML systems is not very unlike developing good software. Just as developing good software requires mastering not only programming theory, tools, and design patterns, but also the process of software development itself, building a good ML system entails familiarity with the ML application lifecycle. In this talk, we will discuss the various stages of ML application life cycle - proble… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
|
Build an enterprise grade data labelling pipeline to scale your ML/AI pipelinesIn Software 2.0, Data is code. A mindful approach to your data annotation pipeline and practices is critical to the outcomes of your ML algorithms. If not done right, your ability to scale this pipeline can often prove to be a major blocker to productionization. more
Session type: Full talk of 40 mins
|
Open Source Tools and Archive for Tackling Misinformation on ChatApps in IndiaTattle is a civic tech project in India that is creating an archive of content circulated on WhatsApp and other chat apps, and building open source tools to navigate this archive. Such an archive is useful for research on information networks as well as for increasing the efficiency and reach of fact checking efforts. One of Tattle’s goals is opening the archive, even if in a limited scope, to th… more
Session type: Short talk of 20 mins
|
Gendered Biases in Artificial IntelligenceMy talk will attempt to bust the myth of “objectivity” or “neutrality” of Artificial Intelligence (AI) technologies by highlighting gendered biases in AI and how they arise. To substantiate, I will focus on Virtual Assistants / “Bots” by presenting my primary research on a comparative analysis of the responses of smartphone-based Virtual Assistants (such as Siri and Alexa) to user queries on gend… more
Section: Full talk
Technical level: Intermediate
Session type: Lecture
Session type: Full talk of 40 mins
|
How we applied sampling algorithms to extract meaning from data (@ Belong.co)A lot of unsupervised learning algorithms work by inferencing parameters of generative models through Monte Carlo techniques. In this talk, we will go into details of the underlying inference algorithms that use sampling techniques and then proceed step-by-step applying it to couple of real world problems, particularly some of our work at Belong that we recently published at ICDAR’19. The attende… more
Session type: Full talk of 40 mins
|
Efficient Machine Translation for low resource languages using TransformersTransformer is the first transduction model relying entirely on self-attention to compute representations of its input and output without using sequence aligned RNNs or convolution. Transformers were recently used by OpenAI in their language models, and also used recently by DeepMind for AlphaStar, their program to defeat a top professional Starcraft player. more
Session type: Full talk of 40 mins
|
Hosted by