Data is the new oil fueling Machine learning adoption but Businesses are discovering #AI is no silver bullet

Data is the new oil. However, unlike oil, as data scarcity is becoming less of a problem, processing costs are skyrocketing. The business world is waking up to the fact that while the cost of computing keeps getting cheaper all the time. The cost of training machine learning models is outpacing the compute cost drop.

Moreover,  business are finding challenging to adopt #ai, and the economist report numbers are showing how often #machinelearning projects in the real business world fail :

• Seven out of ten said their #ai projects had generated little impact so far.
• Two-fifths of those with “significant investments” in ai had yet to report any benefits at all.

Companies are finding that #machinelearning is not the promised silver bullet. The non-tech company are discovering what tech companies had to learn the hard way: that they are no Google, Facebook, ...

To successfully deploy an AI/ML/DL project you need: a vast amount of data, skilled employee, solid engineering practice, access to infrastructure and last but not least, a clear understanding of the business problem.

I have a false hope that corporation will abandon the silver bullet thinking, but I would settle for avoiding another #ai winter cycle.

[Links of the Day] 19/03/2020 : Directed Acyclic Graph structure estimation, Groovy Linter, AI hierarchy of needs

• DAGs with NO TEARS : NIPS 2018 paper that demonstrate a novel way to Estimate the structure of directed acyclic graphs. Bonus point for code in github ! [arxiv]
• groovyfmt: I wish I knew about this one a long time ago. All those Jenkins File errors and debugging session I could have avoided. Well, let's add it to my default list of linter to run with every job.
• AI hierarchy of needs : neet representation of what is needed to deliver an AI project and how much effort and information is required as you progress throughout the hierarchy.

[Links of the Day] 07/01/2020: 2019 AI index report, Essential Guide to electronics in Shenzhen, Tech Lead Expectations for engineering projects

AI index report 2019 : a little bit generic but still a good refresh of where we are and where we may be going.
Essential Guide to Electronics in Shenzhen : Huaqiangbei electronics market in Shenzhen is a must-visit for any tech aficionado. On top of the diversity of things, you can find you literally have access to a swarm of manufacturer and engineering group that can build whatever you need or can think of under one roof.
Tech Lead Expectations for Engineering Projects : an awesome overview of the tech lead by Gergely Orosz @uber. It provides a good overview and guidance of what this role consists of, and what the expectations are.

[Links of the Day] 19/12/2019 : Machine learning at arXiv, Netflix human centered machine learning infrastructure management library, Filesystems are still not fully SSD aware

• arXiv Machine Learning Classification Guide : how does ArXiv classify papers automatically with machine learning and what they plan to do with it in the future.
• Metaflow : Netflix open source it's human-centred python library for managing machine learning infrastructure. It can user PyTorch, Tensorflow and Scikit. [website]
• Evaluating File System Reliability on Solid State Drives : Your filesystem needs to understand the underlying hardware in order to guarantee reliability and security of the data stored. Nowadays the majority of storage solution relies on SSD but the authors discovered that many of the common filesystems are not fully SSD aware in their operations. They demonstrated that in 16% of the case faults resulted in irrecoverable failures.

[Links of the Day] 25/10/2018 : Distributed AI framework, Reverse proxy API gateway, AirBnB Change data capture service

• Ray : A Distributed Framework for Emerging AI Applications [Github]
• annon.api : Configurable API gateway that acts as a reverse proxy with a plugin system.
• SpinnalTap : Change Data Capture (CDC) service capable of detecting data mutations with low-latency across different data sources, and propagating them as standardized events to downstream consumers.

[Links of the Day] 26/10/2017 : Algebra for #AI , Live presentation tool, The (nonexistent) CEO effect

• All the Linear Algebra You Need for AI : well not really, but it is a good start
• doitlive : if you ever had to do a live demo/presentation. You know how dreadful it can be. Stuff can go wrong really fast. This tool allows you to remove some of the hassles of this type of exercise. It reads a file of shell commands and replays the commands in fake terminal sessions. 
• How large must the CEO effect be to rule out chance? : turn out that most CEO are completely interchangeable as that most CEO has little effect on company performance. 

[Links of the Day] 05/05/2017 : Machine learning Cheet Sheet, Sarcasm corpus for your #AI and hardening your architecture

• Machine Learning Cheat Sheet : Ok it's more than one pager, but it contains pretty much all the essential machine learning info.
• A Large Self-Annotated Corpus for Sarcasm : If you want to make your chatbot even more annoying .. This is what you need. 
• Resilient Functional Service Design : Hardening your distributed architecture the right way

[Links of the Day] 28/04/2017 : Bitcoin Antbleed , Social Networks Rumors, HPC & AI trends

• Increasing the Flow of Rumors in Social Networks by Spreading Groups : Looks like by fragmenting groups rumours flow more easily in social networks. To a certain extent, this mimics real life as by isolating and fragmenting group it becomes easier to spread gossip due to the difficulty by an individual in each group to check the validity of the information within its neighbourhood.
• HPC & AI Technology Trends : Dr Eng Lim Goh of HPE talks about the trend in HPC and AI.
• Antbleed : Apparently, BITMAIN,  the ASIC system provider of up to 70% of bitcoins miner embedded a backdoor that can disable or compromise remotely its hardware. The funny aspect is that it can potentially allow the company to pass the 51% control of bitcoin miner network, and hence allow it to rewrite the whole blockchain. The 51% threshold has always been considered as a theoretical threat that was not attainable in real circumstance. Well, guess what, it's not theoretical anymore.

[Links of the day] 19/10/2016 : #AI hard problems, Dark Silicon & Reliability , Transport Layer Dev Kit

• Applied AI hard problems : current and future AI hard problem, the interesting bit is the "emergent" behavior aspect that computer scientist are trying to achieve. Where AI is not tailored for a specific problem by adapt to the environment it encounter. 
• Dark silicon & Hardware Reliability : the authors look at the impact of the dark silicon approach ( when not all component are turned on when the system is up) and how to leverage the "dark" ratio to maximise lifespan of hardware. [slides]
• TLDK : project lead by Intel within the fd.io framework. It is trying to adresse the lack of high level ( as in layer 4 ) packet processing capabilities. The project aim at delivering UDP/TCP etc.. packet processing on top of vector packet processing of FD.io (which can works on top of DPDK). By doing so Intel will be able to finally have a comprehensive framework which will enable DPDK based solution to flourish beyond the pure networking stack (NFV) solution.

[Links of the day] 19/09/2016 : #AI bias, Incremental consistency , Customizable datacenter

• Stuck in a Pattern : as predictive policing tools are being widely adopted in corporation and public organisation. There is little transparency as how these systems have been configured. It seems that the current set of software designed and deployed may reinforce discrimination and inequality under a veil of marketing publicizing intelligent solution. 
• Incremental consistency guarantees : The authors propose a system that instead of providing a single "hard" consistent answer to a query a system that will provide multiple reply with incremental consistency guarantee albeit with incremental latency cost. This allow system to make decision based on their consistency requirement as well as performance needs. This is interesting as it would allow some application to take decision based on consistent enough information while being able to revise their decision if needed once receiving a higher level of consistency response.
• Customizable Computing at Datacenter Scale : NAS 16 keynote , it seems that HPC and exascale system are slowly converging toward an hybrid model with heterogeneous resources, FPGA, GPGU , CPU , etc.. 

[Links of the day] 14/09/2016 : Ethic in AI , Survey of fully homomorphic encryption, RDMA over Ethernet at scale at Microsoft

• Ethical Preference-Based Decision SupportSystems : when AI and other autonomous agent start to be more ubiquitous in the human environment. As the decision of these systems will start to have a greater impact on our daily life, trust will need to be build and to achieve that these system will need that they are perceived to act in a moral and ethical way. 
• A brief survey of Fully Homomorphic Encryption, computing on encrypted data : fully homomorphic encryption allow you to manipulate encripted data without decrypting it. This is great for database and other systems as it allow service to modify and update information without the need to know its content. Effectively partitioning operation from knowledge. However this comes at a cost (but its going down). Might finally end up with the security pipe dream where the data is immediately encrypted and is only manipulated in this form until it is finally consumed.
• RDMA over Commodity Ethernet at Scale : It is interesting to see that RDMA start to slowly permeate hyper-scale data-center. However it is even more interesting to see that Microsoft decided to go for the RoCE version of it instead of infiniband. It make sens as there was a lot of investment in scaling the Ethernet for their cloud infrastructure and allow a lot of reuse and collocate normal and RDMA traffic on a single underlying fabric.

Links of the day 02/12/2015: fast linux perf analysis, Path to AI, Datacenter transport

• Linux Performance Analysis in 60,000 Milliseconds : fast minimal performance analysis approach allowing to quickly narrow down where the issue might come from. 
• Path to general AI : interesting essay on AI and why the current path for will not allow the emergence of a true intelligence as human understand it.
• pHost : Distributed Near-Optimal Datacenter Transport Over Commodity Network Fabric. This is an improvement over the previous Fastpass transport as it allows end-hosts to directly make scheduling decisions, thus avoiding the overheads of centralized scheduler architecture

Links of the day 02 - 02 - 2015

Today's links 02/02/2015: #AI , #Debian #Openstack Image, Metrics Driven prioritization and search engine

• Awesome Artificial Intelligence : A curated list of Artificial Intelligence (AI) courses, books, video lectures and papers 
• Openstack Debian Image : the official OpenStack Debian image is now generated at the same time as the official Debian CD ISO images. If you are a cloud user, if you use OpenStack on a private cloud, or if you are a public cloud operator, then you may want to download the weekly build of the OpenStack image from here: http://cdimage.debian.org/cdimage/openstack/testing/
• Metrics-Driven Prioritization : integrate business metrics and probabilistic modeling into prioritization process. 
• Hound : an extremely fast source code search engine.