Ray­mond Ram­char­i­tar

The term "think tank" de­scribes an in­sti­tu­tion which pays peo­ple to think, cre­ate, and pur­sue so­cial and sci­en­tif­ic re­search for their own sake. It has been a large­ly met­ro­pol­i­tan phe­nom­e­non till now, but has fi­nal­ly reached the Caribbean out­side the acad­eme in TT­LAB, thanks to Prof Patrick Ho­sein, a MIT-trained com­put­er sci­en­tist who works at the Uni­ver­si­ty of the West In­dies (UWI). He is al­so the hold­er of sev­er­al patents in cel­lu­lar phone and In­ter­net tech­nol­o­gy, and is the ANSA Caribbean Awards Sci­ence & Tech­nol­o­gy Lau­re­ate 2015.

One of Prof Ho­sein's peeves about re­gion­al acad­e­mia and de­vel­op­ment agen­das is that there has been much talk but lit­tle re­al un­der­stand­ing about in­no­va­tion and its re­quire­ments. What we think is in­no­va­tion, he says, is usu­al­ly just im­port­ing and mod­i­fy­ing for­eign tech­nol­o­gy. As a re­sult, many op­por­tu­ni­ties for in­no­va­tion are lost, or sim­ply with­er on the lo­cal vine. One ex­am­ple which was cru­cial to TT­LAB in­volved a stu­dent at UWI, St Au­gus­tine, where Prof Ho­sein teach­es com­put­er sci­ence.

The UWI stu­dent, Ste­fan Ho­sein (no re­la­tion to Prof Ho­sein), had been to Nasa (Na­tion­al Aero­nau­tics and Space Ad­min­is­tra­tion in the US) on an in­tern­ship/schol­ar­ship, thanks to the T&T Na­tion­al In­sti­tute of High­er Ed­u­ca­tion, Re­search, Sci­ence & Tech­nol­o­gy (Ni­herst). One of the con­di­tions of the schol­ar­ship was that he had to re­turn home to serve in a gov­ern­ment in­sti­tu­tion.

When he re­turned to the UWI, said Prof Ho­sein, he was giv­en rel­a­tive­ly me­nial tasks to per­form, like web de­vel­op­ment. He was ca­pa­ble of much more, and Prof Ho­sein de­cid­ed to en­gage in an ex­per­i­ment: pay him a stipend, al­low him to pur­sue re­search, and see if he could come up with a pub­lish­able pa­per with­in three months.

Ste­fan went on to pub­lish four pa­pers in peer-re­viewed in­ter­na­tion­al con­fer­ences dur­ing a one-year pe­ri­od. These were based on his re­search at TT­LAB but he al­so col­lab­o­rat­ed with Nasa on oth­er pub­li­ca­tions dur­ing this pe­ri­od. This was TT­LAB's first fel­low. "TT­LAB has helped me tremen­dous­ly," Ste­fan said.

"I was able to pro­duce re­search (some­thing that is rare for a per­son with­out a post­grad­u­ate de­gree in Trinidad) which helped me to se­cure ac­cep­tance to The Uni­ver­si­ty of Cam­bridge for my post­grad­u­ate de­gree. Ad­di­tion­al­ly, I was able to at­tend one of the pre­mier con­fer­ences in my re­search area, this al­lowed me to net­work with those who are top in the field and get a bet­ter un­der­stand­ing of what the state-of-the-art re­search looks like."

TT­LAB is fund­ed en­tire­ly by Prof Ho­sein's com­pa­ny, the T&T Net­work In­for­ma­tion Cen­tre (TTNIC). It seeks and at­tracts fel­lows who range from un­der­grad­u­ates pur­su­ing high­er de­grees to sci­en­tists who can­not do re­search in their work­places, of­fers them fund­ing, and lets them pur­sue projects they want to. The re­sults have been spec­tac­u­lar. From its one stu­dent/re­searcher, a lit­tle over a year ago the project now has 19 fel­lows and re­searchers. One is an­oth­er Nasa in­tern, In­za­mam Ra­haman, who was in­tro­duced to Prof Ho­sein by his first pro­tege.

Ste­fan's work is most­ly in Ma­chine Learn­ing, but Ra­haman's work is in math­e­mat­i­cal op­ti­mi­sa­tion, di­rect­ed at us­ing so­cial net­works, like Twit­ter and Face­book. Ad­di­tion­al­ly, he is pur­su­ing re­search (with Shan­ta Sukhoo) on "how we can char­ac­terise the re­la­tion­ship be­tween col­lab­o­ra­tion and in­no­va­tion en route to build­ing rec­om­men­da­tion sys­tems for team build­ing". He is al­so in­ter­est­ed in the In­ter­net of Things and its po­ten­tial in de­vel­op­ment.

About TT­LAB

These are just two re­searchers and fields, there are many more. On its web­site (http://lab.tt) TT­LAB lists the va­ri­ety of its achieve­ments: 15 pub­li­ca­tions and con­fer­ence pa­pers, which is re­mark­able by any stan­dard for a year's en­deav­our. The top­ics ad­dressed range from op­ti­mis­ing elec­tric­i­ty grids, gaug­ing emo­tion­al re­spons­es from so­cial me­dia ("Sen­ti­ment Clas­si­fi­ca­tion from Twit­ter Feeds"), an app for the de­tec­tion of dan­ger­ous dri­ving, and the ge­net­ics of non-com­mu­ni­ca­ble dis­eases in the re­gion.

To spur its re­searchers, any mem­ber of TT­LAB, who has an ac­cept­ed con­fer­ence pa­per, is pro­vid­ed with a trav­el grant to at­tend and present in at least one con­fer­ence per year. TTNIC al­so funds full fel­low­ships for two PhD and two MPhil stu­dents. This year, three full schol­ar­ships have been award­ed to three 2016 first-class ho­n­ours stu­dents to pur­sue their MSc de­gree in Com­put­er Sci­ence at the UWI.

Away from its re­search con­cerns is a cul­tur­al one: TT­LAB is al­so a con­crete re­sponse to one of Prof Ho­sein's bug­bears: that tech­nol­o­gy and ex­per­tise are im­port­ed when they are avail­able at the same qual­i­ty for a cheap­er price lo­cal­ly. This was al­so a con­stant com­plaint of the late Prof Dave Chadee, a world fa­mous en­to­mol­o­gist, con­tract­ed by in­ter­na­tion­al agen­cies, who was all but ig­nored in T&T.

"A lot of lo­cal com­pa­nies use Huawei tech­nol­o­gy," said Prof Ho­sein. "I worked at Huawei and have about 15 patents in my name which are in­te­grat­ed in­to their prod­ucts. So my work is be­ing used, but no one will ac­cept my help when I of­fer it lo­cal­ly."

TTNIC man­ages the reg­is­tra­tion of ".tt" do­main names. It al­so pro­vides free do­main names (un­der edu.tt) to all ed­u­ca­tion­al in­sti­tu­tions and of­fers free web­site host­ing to those that are reg­is­tered with the Gov­ern­ment. In ad­di­tion, it pro­vides spon­sor­ships to many lo­cal ICT re­lat­ed events.

Do­main name poli­cies are de­cid­ed by a mul­ti-stake­hold­er ad­vi­so­ry group, the TTMAG (see mag.tt). TTNIC is al­so ac­tive in Open Da­ta and man­ages the sites da­ta.tt and maps.tt. Its funds are lim­it­ed, but be­ing en­tire­ly self-fund­ed means it doesn't have to an­swer to im­pa­tient busi­ness­es or wait on bu­reau­crat­ic fun­ders.

And here aris­es a ques­tion of why TT­LAB's es­tab­lish­ment had to take so long, and await a sin­gle in­di­vid­ual to do it. T&T has three uni­ver­si­ties and (up till 2016) uni­ver­sal free ter­tiary ed­u­ca­tion, yet, there has been no com­pa­ra­ble en­deav­our.

Cul­tur­al prob­lem hin­der­ingre­search and de­vel­op­ment

In­com­ing UWI, St Au­gus­tine, Pro Vice Chan­cel­lor Prof Bri­an Copeland dis­agrees about the dis­par­i­ty in out­put. There has been much re­search and in­no­va­tion at the UWI, he said.

"UWI has a range of units and cen­tres that con­duct re­search on a va­ri­ety of top­ics. The Seis­mic Unit, for ex­am­ple. At­ti­tudes to the full range of the S&T spec­trum (blue sky to ap­plied) have been chang­ing over the past few years. The work of St Clair King's Re­al Time Sys­tems Group, the work of those who gen­er­at­ed patents over the years (Prof Ram­sey Sanders and oth­ers, in­clud­ing my­self), the work at the Co­coa Re­search Cen­tre which has just seen the sign­ing of MOUs with ex­ter­nal com­mer­cial en­ti­ties, and the work at the Steel­pan Ini­tia­tives Project for de­vel­op­ing the GPan and PHI for com­mer­cial pro­duc­tion."

As to out­put and qual­i­ty of work, said Prof Copeland: "My most re­cent ex­pe­ri­ence in En­gi­neer­ing was of stu­dents with mod­est po­ten­tial, as mea­sured by the course­work GPA, ris­ing to the chal­lenge of fi­nal year projects and, in some cas­es, sub­mit­ting award-win­ing pa­pers in­ter­na­tion­al­ly." The UWI, he con­tin­ues, has al­so de­signed apps for lo­cal fish­er­folk, its Med­ical Sci­ences staff is pro­lif­ic in their work, and its Dean is ranked in the top 100 "most in­flu­en­tial pa­pers on Chron­ic Ob­struc­tive Pul­monary Dis­ease (COPD)."

There is no doubt the UWI has had a strong im­pact re­gion­al­ly and pro­duces high-qual­i­ty re­search. But two com­plaints about the UWI are its dis­con­nect­ed­ness from the so­ci­eties it serves, and its in­ter­nal pol­i­tics. Prof Copeland ac­knowl­edges this. Prof Copeland said: "The con­di­tions at UWI are in­deed less than ide­al and the staff, who have ral­lied, de­spite this, are to be cel­e­brat­ed. Patrick is an ab­solute gem pri­mar­i­ly be­cause of his ex­pe­ri­ence at Bell Labs (we do not get many of those) and his pro­lif­ic patent [record].

"He has at­tract­ed stu­dents to TT­LAB be­cause of this. He has seen the mech­a­nism of con­cept to com­mer­cial re­al­i­ty first hand at Bell. It costs to get peo­ple to that lev­el of achieve­ment...you should ask him to es­ti­mate the cost to Bell Labs for his patents. This coun­try grudg­ing­ly in­vests 0.04 per cent of its GDP in RDI.

The most di­rect RDI sub­ven­tion by gov­ern­ment of Trinidad & To­ba­go is a pal­try $5 mil­lion a year. In­dus­try and com­merce pro­vide much less than that. The one or two gen­uine and valid at­tempts to pro­duce new prod­ucts with in­ter­na­tion­al mar­ket po­ten­tial have been all but slaugh­tered. There is a cul­tur­al prob­lem in the way of progress."

All this notwith­stand­ing, Joyce­lyn Lee Young, a for­mer act­ing pres­i­dent of Ni­herst, said TT­LAB's "R&D pro­duc­tiv­i­ty is phe­nom­e­nal. The Re­al Time Sys­tems Group run by Prof St Clair King in the 1990s was close but I don't re­call that it achieved this much in a one-year time­frame. TT­LAB would have achieved a first to my mind with Ste­fan Ho­sein do­ing four pa­pers at con­fer­ences and Imza­man Ra­haman hav­ing three, both with just a BSc."

An im­por­tant com­po­nent of TT­LAB's suc­cess, con­tin­ued Ms Lee Young, is its per­son­al­i­ty, and the per­son­al­i­ty of its founder. She said: "I am amazed to see how Patrick/TT­LAB took two of our Nasa in­terns to new heights in R&D–from the con­cep­tu­al­i­sa­tion stage to ex­e­cu­tion, the ro­bust method­ol­o­gy, and high-lev­el prob­lem solv­ing un­der­tak­en. Both in­terns had just a BSc but their work was at the lev­el of post-grad stu­dents at MIT or Stan­ford. To me it showed what a men­tor who has worked on the world stage could do with bright young minds in a lo­cal set­ting. He is pas­sion­ate about de­vel­op­ing young re­searchers, in­spir­ing them to be­lieve in them­selves and guid­ing them to­wards achiev­ing high qual­i­ty R&D."

Apro­pos of Prof Copeland's iden­ti­fi­ca­tion of the "cul­tur­al prob­lem" hin­der­ing re­search and de­vel­op­ment, an im­por­tant part of TT­LAB's char­ac­ter is its up­end­ing of con­ven­tion­al de­vel­op­ment ap­proach­es. "There's a lot of em­pha­sis lo­cal­ly on build­ing large build­ings, and then there's no re­sources left for re­search," says Prof Ho­sein. TT­LAB is the op­po­site. It has no head­quar­ters or bu­reau­cra­cy.

The com­pa­ny sim­ply se­lects the best stu­dents (or in­vites ap­pli­cants it be­lieves are gift­ed) gives them mon­ey, and lets them do what they want. He treats them the way that he likes to be treat­ed which is to be giv­en au­ton­o­my to ex­plore and have com­plete time flex­i­bil­i­ty (they can work from home if they wish and, like him, many pre­fer work­ing at nights). Un­for­tu­nate­ly he can­not pay as much as he thinks they de­serve to be paid but the Fel­lows ac­tu­al­ly re­ceive a larg­er stipend than that pro­vid­ed by UWI schol­ar­ships.

Lee Young agrees: "In Patrick's field T&T can be­come a glob­al play­er in R&D–T&T does not need ex­pen­sive huge labs. We are deal­ing with brains–rea­son, log­ic, com­pu­ta­tion, and soft­ware skills. We can de­vel­op and har­vest this re­source in a short space of time–com­pared to R&D in oth­er dis­ci­plines."

This ap­proach is par­tic­u­lar­ly use­ful, says Prof Ho­sein, to those sci­en­tists and en­gi­neers who work in day jobs, but are not al­lowed to do the kind of re­search they want to do. With a lit­tle ex­tra mon­ey, they can take time off from work, and spend it on the things that mat­ter more. He be­lieves that a "re­verse-in­tern" pro­gram where­by bright en­gi­neers/sci­en­tists in in­dus­try get to per­form re­search un­der the su­per­vi­sion of UWI Fac­ul­ty might be fea­si­ble and all three par­ties will ben­e­fit.

In­deed, he says, it was his prize from the ANSA Caribbean Awards that al­lowed him to take time off from his teach­ing at UWI to pur­sue his own re­search and help fund his first ex­per­i­men­tal stu­dent.

An­oth­er in­no­v­a­tive prac­tice is en­cour­ag­ing stu­dents to think of what they would like to pur­sue, as well as en­cour­ag­ing top­ics of lo­cal use and in­ter­est but with glob­al scope. "To pub­lish in­ter­na­tion­al­ly, you have to be more broad­ly fo­cused than just Trinidad & To­ba­go," says Prof Ho­sein, "but there are things here that qual­i­fy." One of these things is elec­tric­i­ty grid ef­fi­cien­cy.

TT­LAB (through a RDI Smart Grid project lead by Dr Shar­ma at UWI) ap­proached T&TEC for da­ta from their me­ters which they analysed and de­vel­oped an al­go­rithm and an ap­pli­ca­tion which could pro­vide sig­nif­i­cant sav­ings if ap­plied since com­pa­ra­ble for­eign soft­ware is ex­pen­sive. The project, un­for­tu­nate­ly, has been frozen.

But en­gi­neer­ing is not the on­ly area that can ben­e­fit. The UWI (and every uni­ver­si­ty li­brary) main­tains sub­scrip­tions to in­ter­na­tion­al jour­nals which, must be paid for in US cur­ren­cy. Analysing the da­ta on how UWI mem­bers ac­cess the jour­nals, TT­LAB de­rived an op­ti­miza­tion al­go­rithm which rec­om­mends which jour­nals should be sub­scribed to, and which sub­scrip­tions can be halt­ed, and in­di­vid­ual ar­ti­cles pur­chased in­stead. This would mean no loss of ac­cess for stu­dents and fac­ul­ty, but mil­lions of dol­lars can po­ten­tial­ly be saved by the uni­ver­si­ty.

The biggest prob­lem in putting this knowl­edge to work, says Prof Ho­sein is lo­cal in­er­tia. "We have tried to talk to lo­cal com­pa­nies, but they on­ly see us as app de­vel­op­ers, and any­one with com­put­er knowl­edge can de­vel­op an app. Most don't see the val­ue in our ex­per­tise in Ma­chine Learn­ing, Op­ti­miza­tion, De­ci­sion Sup­port Sys­tems and so on. Just as in the above cas­es we can use this ex­per­tise to help them achieve sig­nif­i­cant sav­ings.

"And even when the re­turns are im­me­di­ate, as in the T&TEC re­search to im­prove grid qual­i­ty, bu­reau­cra­cy freezes the process. In fact, says Ho­sein, many lo­cal com­pa­nies tend to pre­fer to get for­eign con­sul­tants. If TT­LAB can con­tin­ue to grow, this could change, as "we would like to go be­yond just pub­lish­ing, and start fil­ing patents and de­vel­op­ing lo­cal prod­ucts to show we can ben­e­fit lo­cal in­dus­try."

TT­LAB's founder wants to keep it fo­cused on re­search but ad­mits that some de­vel­op­ment may be re­quired in the fu­ture to help pay the bills. "The pre­ferred mod­el is sim­ple," he said. "A com­pa­ny ad­mits they need help, pro­vides us with the in­for­ma­tion (es­pe­cial­ly da­ta) that we would need, we pro­vide a so­lu­tion and pos­si­bly a pro­to­type of the so­lu­tion and we move on. Of course dur­ing the process we hope to pub­lish but we will nev­er dis­close any con­fi­den­tial or com­pa­ny pro­pri­etary in­for­ma­tion."

The ob­ject and means to de­vel­op might be sci­ence, but the spir­it that al­lows it to ex­ist must be cul­tur­al.

�2 Ar­ti­cle cour­tesy ANSA Caribbean Awards