Technology.

Technology Overview

ForwardLane employs machine learning, machine intelligence, and statistical optimization methods to synthesize valuable insights.

Proprietary IP Development

We use leading scientific open source platforms to develop our core technology focused on natural language processing of financial domain data with machine learning, deep neural networks,  and knowledge graphs.

Commercial Partners

This enables us to integrate fundamental research, thematic views from the CIO, product documentation, in depth news, commentary and chatter with market data as reflected in asset prices, time series data, macroeconomic factors and through rich, evolving knowledge graphs.

We have deeply integrated quantitative factor models, risk analytics, scenario and backtesting capabilities.

  • SYNTHESIS ENGINE
  • INSIGHTS ENGINE
  • Conversational Engine
  • PAAS-CLOUD FOUNDRY

Our platform services are underpinned by cutting-edge infrastructure services designed to be ultra-fast, petabyte scale and distributed.

Our solutions meet modern bank infrastructure requirements with on-premise, hybrid and cloud deployments, an enterprise integration layer that takes less than a week to integrate with thousands of enterprise applications and 256-bit AES encryption in flight and at rest.

  • DeepLake - Data LAKE
  • Blackwidow - Graph
  • HyperFlex - Distributed Data Engine
  • Snowflake - Search
Platform Services

ForwardLane’s Platform Microservices can be dynamically combined in realtime through our API Gateway to deliver scalable solutions for our clients.

SYNTHESIS ENGINE

We process large volumes of structured and unstructured data on a highly scalable basis.

Machine learning and natural language processing of unstructured real-time market data, news, and research.

Leveraging small and big data processing — achieving accuracy of 86.7% on hold-out test financial domain data with and error of +/- 2%. Training data accuracy is in the order of 92.7% for financial Q&A datasets.

Conversational Engine

The ForwardLane Intelligence engine can craft better conversations compared to traditional rules based conversational bots.    We use statistical inference and predictive analytics on large data sets, coupled with conversational memory and knowledge graph a to deliver a more accurate and personalized conversational experience.

INSIGHTS ENGINE

Our engine can integrate existing analytical models your firm has built and deliver deeper insights to personalize the client engagement interactions. We enrich data from the graph, integrate demographics, trading history, predictive models:
— understand and visualize relationships
— from neural cloud
— demographic, interests

PAAS-CLOUD FOUNDRY

Hardened, Production-ready infrastructure —  using Industry-standard platform services trusted by IBM, GE, HP Enterprise and SAP, ForwardLane’s APIs and Microservices are managed for deployment at scale in enterprise production environments. Full flexibility for deployment via secure cloud, hybrid or on-premise.

Infrastructure Services

ForwardLane’s Infrastructure services combined big data processing capabilities with distributed graph and search to do the heavy-lifting so that our clients don’t have to. This backend powerhouse ensures we deliver the smartest possible solutions to our clients.

DeepLake - Data LAKE

Our data is enriched with data harvested from neural networks and machine learning algorithms.

We process large volumes of structured and unstructured data using NoSQL data lake for ingesting and munging sell-side research, real-time news, asset prices, time series data, product data on 35,000 companies in 450 markets with equity, commodities, fixed income, foreign exchange and money market data as well as mutual funds, ETFs, hedge funds and indices.

HyperFlex - Distributed Data Engine

Peta-byte scale, ultra-fast distributed data store leveraging technologies based on Hbase.

Processing billions of data points at terabyte scale in seconds enables ForwardLane to make sense of data fast.

Blackwidow - Graph

We deliver unique insights into complex inter-relationships across markets, events, clients, portfolios and other factors leveraging knowledge graph representations, visualization and industry leading neural networks.

As members of the Enterprise Data Management Council, we have integrated FIBO, the Financial Industry Business Ontology standard, with our proprietary ontologies which are peer-reviewed by FINRA-licensed financial experts with over 8000 financial terms.

Snowflake - Search

Simple, efficient, smart — Proprietary search algorithms bring together both structured and unstructured search results tapping into concepts and their relationships by traversing knowledge graphs and federated, distributed data stores.

How we use machine learning

ForwardLane employs a variety of machine intelligence and learning - primarily natural language processing, deep learning, speech, some vision and bayesian optimization and techniques focused on recommenders and predictive analytics.

Synthesis Engine Overview

We conduct primary research in generative conversation, character-based learning from scratch, very deep convolutional neural networks, and very large classification and Q&A systems.

High-level ForwardLane Data Pipeline

Semi-supervised learning

Semantic analysis

Clustering

Noise Processing

Sentiment mining

Flexible Domain-specific ontologies

Financial Industry Business Ontology (FIBO) contributor

Dynamic generated ontologies

Logical consistent

Hierachical Deep learning

Hierarchical conv-nets and re-nets with svm

Reinforcement learning

n-dimension knowledge graphs

Topological data analysis

Data fusion

Flexible Ontologies

ForwardLane is a member of the Enterprise Data Management Council and contributors to the peer-reviewed Financial Industry Business Ontology. We employ advanced applications of FIBO, in conjunction with our unique flexible dynamic ontologies.

Using logically consistent definitions, we can continuously improve accuracy and be an accurate source trusted by clients.

ForwardLane Visualization of a Core RDF/OWL Ontology

DeepLake Data Aggregation

The ForwardLane DeepLake data aggregation cloud acquires vast amounts of data on a daily basis so that we can offer clients insights across markets including:

Thomson Reuters

Pricing, fundamentals on 35,000 companies worldwide

Realtime Machine-readable news — Corporate and Global a specific event news

Market Psych Sentiment Indices quantitatively driven from Twitter and Social Media

Sell-side Embargoed Research

Starmine Analytics — quantitative Analyst Revisions Models which uses NLP, text mining and quantitive scoring to predict bullish and bearish signals

Analyst Estimates

Macroeconomic indicators and data

Morningstar

Morningstar Analyst Ratings — including Morningstar Star Ratings

Daily and Historic Data — Open Ended Funds, Closed Ended Funds, Mutual Funds, ETFs, Commodities

Morningstar Indices — Target date, Country, Allocation, Global Equity, Country (Ex-US)

Holdings level data for funds

Hedge fund pricing data

Xignite

Master Reference Data

Global equities pricing for 450 exchanges

Global bond data

FX currency pairs

FCA FSA Finra

Department of Labor and Fiduciary standards

MIFID II

Quovo

Master Reference Data

Global equities pricing for 450 exchanges

Global bond data

FX currency pairs

Soc

Twitter

Linkedin

Demographic Data

Demographics, Interests, income brackets, post code, house value