23 January 2017

Robot Dreams

One might wonder what the benefits are of enabling a robot to dream. However, as it turns out dreaming has significant health benefits also for humans. In humans, dreams have a complex way of consolidating memories and processing emotions. Dreams are inexplicably linked to the mind and can occur during the different phases of sleep or in form of day-dreaming. Respectively, dreams can be valuable for robots to consolidate information overload as well as form complex emotions. They could also help visually evaluate the health performance of a robot through imagery interpretation. Dreams are also a form of conscious experience. Dreams also provide a way to design consciousness in robots that may further provide ways of advancement in emotional experiences and the way they view the world. Dreams are often a simulated experience. They may even add further advancement to the self-awareness of a robot from personal insight, understanding, and originality. And, even further into autobiographical memories. Altered state of consciousness is a powerful experience.

Seven Types of Intelligent Robots

Robots are slowly but surely going to take over the world to some degree with their specialized skills. However, it will be useful to categorize the different forms of human intelligence that can be replicated for robotics. It is understood that every human has a small part of the seven types of intelligence within them. However, over time one area becomes dominant through personal development. The following types are listed below.

  • Linguistic Intelligence - expressive and heightened sense of understanding
  • Logic Intelligence - insightful
  • Kinesthetic Intelligence - heightened sense of space, depth, distance and size
  • Spatial Intelligence - heightened sense of creativity
  • Musical Intelligence - absolute pitch
  • Interpersonal Intelligence - power to influence
  • Intrapersonal Intelligence - deeply connected

Sometimes, an eighth intelligence is also included in form of naturalist which expresses the sensitivity towards nature and their surroundings. However, it is questionable whether this is more an interest rather than intelligence.

Intelligent Machines can so far be split into four core areas of Artificial Intelligence:

  • Reactive Machines - representations are domain specific, no concept of world, memory or past
  • Limited Memory - representation of the world in the past in order to adapt
  • Theory of Mind - representations about the world and about other agents and entities including thoughts and emotions
  • Self-Awareness - representations about self

Blending human intelligence as well as understanding of context from the various representations in memory could add a significant boost to Artificial Intelligence. A further step in this process may extend into the wired personality traits that incorporate the various unique tendencies and preferences of an individual robot.

20 January 2017

Next Generation Web Search

User, Content, and Context become the focal point towards a better and smarter search engine. Some areas of future searchable intelligence are listed below that could potentially add value to search engines of the future or improvement in existing platforms.

  • Better user experience 'how can we help you' bots for searching the Web
  • More semantically aware searching enrichment for contextual understanding
  • Machine Learning for more automated context learning and content filtering
  • Geolocation sensitive searching
  • More personalized searching
  • Searching by collaborative filtering
  • Mood sensitive searching
  • Ubiquitous content path searching and learning through metadata
  • Accessible searching into deeper web
  • Content search visualization and better queries over distributed web links
  • Searching content on specific question phrases similar to jeopardy
  • Better contextual search coverage of the Web
  • Additive context value to knowledge discovery
  • Search integration across products
  • Search becomes everywhere and all knowing, all seeing experience
  • Richer context cues adding more creative options for targeted audience advertising
  • Searching is swifter and shrewder
  • Search engines are able to contextually understand videos, audio, images, and text
  • Searching with better security of user identity and navigational behavior
  • Searching as a form of a crystal ball
  • Searching becomes more ubiquitous through wearable technology and electrical devices/appliances
  • Searching using more semantically aware and game-theoretic crawlers for indexing the Web

How to Design a Successful Twitter

Why would one want to design another Twitter when it is so unsuccessful and is losing money? To learn from the mistakes of others or to design a better and successful Twitter.

How to Design YouTube

How to Design YouTube Part 1
How to Design YouTube Part 2
YouTube Architecture

9 January 2017

Oreilly Free Books

Business
IoT
Data
Programming
Security
Web Development
WebOps
Design

Dark Patterns

Dark Patterns linger on the web as a tell tale signs of user deception and trickery. Often unnoticed by the user through the interface where they are purposely and intentionally made to confuse. They set on a gamut of actions which are unauthorized by the user. The darkpatterns.org provides further details on examples of deliberately confusing and deceptive user interfaces. There are also shady patterns which push boundaries for user desires but are not deliberately deceptive in practice. They may be construed as misdirections with unclear language which tend to trick a user into doing things that they would otherwise not intentionally do. Many companies are aware of such practices but slither within the boundaries of safe zones. And often signs of such deceptions linger in the use of language, the misdirections, as well as when there are too many things happening on the site. Some ads are also deliberately deceptive which use behavior targeting, follow user's web history, and search patterns. 

8 January 2017

Hortonworks Toolset

  • Falcon
  • Atlas
  • Sqoop
  • Flume
  • Kafka
  • NFS
  • WebHDFS
  • Hadoop
  • Hadoop MapReduce
  • Hadoop HDFS
  • Hadoop YARN
  • Pig
  • Hive
  • HBase
  • Accumulo
  • Phoenix
  • Storm
  • Solr
  • Spark
  • Hawq
  • Zepplin
  • Nifi
  • Ranger
  • Knox
  • Cloudbreak
  • Zookeeper
  • Oozie
  • Slider
  • Tez
  • Metron

Stream Processing Engines



SMACK Stack

S : Scala and Spark (The Engine)
M : Mesos (The Hardware Scheduler)
A : Akka (The Actor Model)
C : Cassandra (The Storage)
K : Kafka (The Message Broker)

A Brief History of Smack
Smack Hands-On
Smack Made Simple
Smack Guide
why is smack stack all rage lately
Smack Slideshare
Smack Personalization

Alternatives for Stream Analytics:
GearPump
Flink